Vercel News

Findings from the $1M Vercel Sandbox challenge

October 14, 2026

In August, we launched the $1M hacker challenge for Vercel Sandbox, inviting security researchers to try and escape Vercel Sandbox. We also coordinated with cybersecurity experts Trail of Bits to review the source code and security architecture. Here is a summary of the challenge: 1,285 reports over two weeks $242,000 in bounties paid out One Linux kernel zeroday vulnerability uncovered No reports showed anyone accessing customer data in Vercel Sandbox. The most severe finding was the defect in the upstream Linux kernel. The kernel zero day The program's most important report arrived on day three. Paulos Yibelo (pwn.ai), a HackerOne researcher working from a standard sandbox with no special access, found a defect in the Linux kernel's networking stack. Paulos discovered that specially crafted packets sent from within the microVM enabled three different attacks: Leak host memory over the network Crash the host machine Execute code on the host Leaking host memory In the first attack, memory from the host, the machine your sandbox shares with other tenants, shows up in packets the guest sends over the internet. This can happen continuously, for as long as the guest keeps sending. This

Amazon S3 joins the Vercel Marketplace

October 14, 2026

You can now create and connect Amazon S3 resources through the Vercel Marketplace without leaving your Vercel project dashboard. The Amazon S3 Marketplace integration automatically adds connection details as environment variables and lets you manage resources alongside your application infrastructure, with links to AWS for deeper configuration. Authentication uses short-lived OIDC credentials without storing static AWS keys. Amazon S3 is also available in v0, which provisions resources, configures connections, and generates code from your prompt. type: embedded-entry-inline id: 7qqJ3BsOar5VmX4ISL8oLp type: embedded-entry-inline id: 6clFScOOxbqin6hI02xHXl Vercel Blob provides public and private storage for application file uploads, with built-in caching and CDN delivery close to your end users. Choose the Amazon S3 integration for existing AWS workflows, analytics on managed Iceberg tables with S3 Tables, or vector storage and search with S3 Vectors. Get started in the Vercel Marketplace. Create a new AWS account through Vercel to receive $100 USD in credits, valid for six months, or link an existing account and manage your plan and usage from your Vercel dashboard. Linked accounts

Nemotron 3.5 Lightning now available on AI Gateway for free

October 14, 2026

Nemotron 3.5 Lightning from NVIDIA is now available for free on AI Gateway as nvidia/nemotron-3.5-lightning-free, served by NVIDIA. Nemotron 3.5 Lightning is an open-weight reasoning and chat model for long-running agents, sub-agent work, chatbots, and RAG systems. The model has 30B parameters with 3B active per token, in a hybrid Mixture-of-Experts design that interleaves Mamba-2 and MoE layers with a few attention layers. It has a 1M token context window and supports tool calling and structured outputs. Note: The provider may retain prompts for model improvement. Call Nemotron 3.5 Lightning through the AI SDK, the Chat Completions, Responses, and Messages APIs, or a coding agent connected to AI Gateway. Use nvidia/nemotron-3.5-lightning-free as the model name: To use it in Claude Code, Codex, Cursor, and more, install the latest Vercel CLI and run setup: Then select nvidia/nemotron-3.5-lightning-free in the agent. See the coding agents guide for details. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, budgets for A

Drama 3 now available on AI Gateway

October 13, 2026

Drama 3 from Fish Audio is now available on AI Gateway, and 80% off for 30 days (through Nov 12). The text-to-speech model is currently in preview and lets you direct voice performances with natural-language instructions in your script, making it useful for voiceovers, character dialogue, and audio storytelling. Describe the emotion and pacing in square brackets, add sounds and pauses where they belong, or wrap selected words in whisper or emphasis tags to change their delivery. Directions can be written in a different language from the spoken text. Drama 3 also supports multi-speaker dialogue, with a separate voice for each character. Use fish-audio/drama-3 as the model name to get started: Try Drama 3 in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Read more

MAI-Image-2.6 and MAI-Image-2.6-Flash now available on AI Gateway

October 13, 2026

MAI-Image-2.6 and MAI-Image-2.6-Flash from Microsoft are now available on AI Gateway. Generate images from text or edit existing images through the AI SDK or the AI CLI, with one API key and no provider account. Both models support photorealistic imagery, text rendering, and targeted edits to products, portraits, and layouts. You can provide up to five reference images to guide edits or combine subjects, with output sizes up to 1536 × 1536. MAI-Image-2.6 (microsoft/mai-image-2.6) for final assets, when image quality matters more than speed. MAI-Image-2.6-Flash (microsoft/mai-image-2.6-flash) for faster generation at a lower cost, when you need many iterations, such as exploring prompts or generating variations. Use microsoft/mai-image-2.6 or microsoft/mai-image-2.6-flash as the model ID. To generate an image, call generateImage with a text prompt: Edit images by passing it as input with an instruction: Try MAI Image 2.6 and MAI Image 2.6 Flash in the model playground, or see the image generation guide. Read more

Repository checkout and Git LFS downloads are now faster

October 13, 2026

Repository checkout and Git LFS downloads are now faster on Vercel, reducing the time spent preparing your project's files before a build. During rollout on build machines with more than two vCPUs: The average time to get project files ready fell by 20%. For LFS-enabled projects, the average time to get project files ready fell by 54%, saving approximately 13 seconds per deployment on this step. The 90th-percentile time to get project files ready fell by 22%. This means 90% of deployments finish this step in that time or less. These improvements apply automatically, with no action required. Read the Git deployment documentation to learn more. Read more

Gemini 4 Argon now available on AI Gateway

October 13, 2026

Gemini 4 Argon from Google is now available on AI Gateway. Gemini 4 Argon is built for coding, agentic work, and complex reasoning. It has a 1M token context window, accepts text, image, PDF, and video input, and supports tool calling. On coding tasks, Gemini 4 Argon builds features, fixes bugs, and researches unfamiliar codebases, carrying longer tasks through multiple steps. The model also reads dense diagrams, spreadsheets, layout-heavy documents, and video, which suits tasks like a financial analysis spread across several spreadsheet tabs. Beyond code, it handles work in finance, health, and law, and hard problems in math and science. Gemini 4 Argon uses thinking levels (low, medium, and high) instead of thinking budgets. Call Gemini 4 Argon through the AI SDK, the Chat Completions, Responses, and Anthropic Messages APIs, or a coding agent connected to AI Gateway. Set the thinking level with reasoning in the AI SDK or reasoning_effort in Chat Completions, and see reasoning for the other controls. Use google/gemini-4-argon as the model name: To use it in Claude Code, Codex, Cursor, and more, install the latest Vercel CLI and run setup: Then select google/gemini-4-argon in the ag

Muse Wonder now available on AI Gateway

October 12, 2026

Muse Wonder from Meta is now available on AI Gateway. The model is designed for coding, long-running agents, and computer use. It can plan multi-step tasks, delegate work to sub-agents, and recover from errors as work progresses. Muse Wonder has a 1M-token context window and supports text, image, and PDF input, tool calling, and structured output. Agents can work with code, screenshots, and documents in the same workflow. You can access the model through AI Gateway with one API key, without creating a separate Meta account. Set AI_GATEWAY_API_KEY in your environment, then use meta/muse-wonder-1.0 as the model name: For coding agents, run npx vercel ai-gateway setup and select meta/muse-wonder-1.0 in your agent's model settings. Try Muse Wonder in the model playground, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own provider key. Read more

Agents can now buy domains with the Vercel CLI

October 9, 2026

Your agent can now use the Vercel CLI to buy a domain. The agent discovers candidates with vercel domains search, checks availability and registrar pricing with vercel domains check and vercel domains price, and initiates the purchase with vercel domains buy. The purchase decision itself stays with you by design. Run non-interactively, vercel domains buy returns a structured, machine-readable error and the suggested next commands, so the agent hands the confirmation back to you, prohibiting it from spending money without your approval. You can confirm the price, the auto-renew choice, and the registrant contact details in an interactive terminal or the dashboard. The CLI ships an agent-facing skill covering the discovery, pricing, and purchase workflow, so agents will follow the full sequence without being told the steps. Read the documentation to get started. Read more

New Pro teams now default to 30-day deployment retention

October 9, 2026

New Pro teams now default to a 30-day deployment retention period. Coding agents have accelerated development, increasing deployment frequency and the amount of stored deployment history. Shorter retention keeps recent history available while limiting older history and its storage costs. You can choose a longer retention period in your team's settings when you need more history for inspection or rollback. Deployment Storage and Functions Storage each cost $0.10 per GB-month. Older deployments may be deleted under your retention policy. Deleted deployments cannot be restored or used for rollback. Review the Deployment Retention Policy and storage optimization guide. Read more

Deployment Storage billing begins for existing Pro teams

October 9, 2026

Deployment Storage and Functions Storage billing will begin for all Pro teams at $0.10 per GB-month. Check your email for details about when billing will start for your team. To keep storage costs low, the deployment retention window also changes to 30 days. Deployments older than 30 days will be deleted starting October 23, unless you opt out in your retention settings before then. Retention settings already at 30 days or shorter are unchanged. Control your Deployment Storage Review retention periods for Pre-Production, Production, Canceled, and Errored deployments with a Deployment Retention Policy. Shorter periods reduce stored history, and deleted deployments cannot be used for rollback. See Deployment Storage and Functions Storage per project on the Usage page. Reduce future deployment size by removing unnecessary build files, moving large assets to Vercel Blob, and shrinking Function bundles. Read the storage optimization guide. Read more

How Rillet ships 3× faster with AI agents on Vercel

October 9, 2026

Rillet on Vercel Tripled shipping rate in three months with eve agents deployed on Vercel Two-person team closed 800+ Linear tickets in its first 4 months Builder access via Enterprise Managed Users, SSO, and Directory Sync Ships customer-requested changes to production in as little as two hours Gives every builder access to models through AI Gateway Rillet is an AI-native enterprise resource planning (ERP) platform. Their AI agents do accounting work inside a real-time general ledger, with human approval and a full audit trail. Rillet serves more than 600 customers working toward a zero-day close, reducing the time needed to close their books at month-end. A product that helps customers close their books in near real time needs a team that can respond quickly to their requests. Anthony Liang heads Rillet’s two-person frontend team, formed only a few months ago, and leads the push to put agents to work across internal workflows, from shipping customer fixes to reviewing code. The team builds agents with eve and runs them on Vercel, using Vercel Connect to give them access to Slack and Linear, while AI Gateway routes their model requests. Standardizing every internal agent on eve An

Microsoft Decision-1 now available on AI Gateway

October 9, 2026

Decision-1 from Microsoft is now available on AI Gateway. It answers typed questions about text with probabilities, choices, and scores. It can be used to classify support requests, select the next step in a workflow, or assess an AI response against a rubric. The model can answer yes/no questions, choose among named categories, or score an input against an ordered set of criteria. Multiple questions can share the same input in one request. Choice and score answers include probabilities for each option, so your application can decide when to act or send an uncertain result for review. Call Decision-1 through the AI SDK decision API, the OpenAI-compatible Decisions API, or the TypeSafe-compatible API. You only need an AI Gateway API key, with no separate Azure account or deployment. To get started, use microsoft/microsoft-decision-1 as the model name in your requests: AI Gateway provides a unified way to access models and track usage and cost. Decision calls appear in logs and count toward budgets alongside other model requests. See the decision documentation for Choice, Score, and Boolean questions. To pick a decision model, browse the available models or read our decision model gu

Liquid AI's d1 is available on AI Gateway

October 9, 2026

Liquid d1 is now available on AI Gateway. d1 is a decision model that evaluates shared state against typed questions for classification, routing, and scoring. It returns structured answers with probabilities without generating text tokens. d1 also has vision support: the model can answer typed questions about images for visual classification, inspection, and scoring. Use liquid/d1 through the AI SDK decision API, the OpenAI-compatible Decisions API, or the TypeSafe-compatible API. These examples ask whether a support agent issued a refund: Image input This example reads a local PNG named square.png and asks which color it shows: AI Gateway provides a unified way to access models and track usage and cost. Decision calls appear in logs and count toward budgets alongside other model requests. See the decision documentation for Choice, Score, and Boolean questions. To pick a decision model, browse the available models or read our decision model guides. Read more

Skip sending request bodies to Routing Middleware

October 8, 2026

You can now skip sending client request bodies to Routing Middleware with skipMiddlewareRequestBody. This reduces incoming Fast Origin Transfer usage for middleware invocations and can improve time to first byte, especially for requests with large bodies. Enable this option only if your Routing Middleware doesn’t read request bodies. Vercel Functions and rewrite targets still receive the body, so you can read and process it there. Set skipMiddlewareRequestBody to true in your project’s JSON configuration file, vercel.json: If you use a TypeScript configuration file, add the same setting to vercel.ts instead. This example assumes @vercel/config is installed: Redeploy your project for the change to take effect. The option defaults to false, so existing projects are unchanged unless you enable it. Read more in the skipMiddlewareRequestBody documentation. Read more

Grok Imagine Video 1.5 Lite on AI Gateway

October 8, 2026

Grok Imagine Video 1.5 Lite from xAI is now available on AI Gateway. This lightweight model turns text prompts and images into video with native audio. The model produces clips from 1 to 15 seconds long in 480p, 720p, and 1080p. Seven aspect ratios cover widescreen, square, and portrait formats, so you can create horizontal scenes or vertical videos from the same generation workflow. To start generating videos, set the model to spacexai/grok-imagine-video-1.5-lite in the AI SDK or in the AI CLI. Chain an image model with Grok Imagine Video 1.5 Lite to create a still and animate it in one flow. You can also try Grok Imagine Video 1.5 Lite directly in the AI Gateway Playground. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own provider key. Read more

Vercel Sandbox creation now returns a ready sandbox

October 8, 2026

A new Vercel Sandbox is now ready to run commands and read or write files when creation completes. Sandbox.create() now resolves only after snapshot restoration completes, so the first command or file operation no longer waits for it. Startup errors also reject Sandbox.create() instead of surfacing during that operation. Snapshot restoration time now counts toward creation instead of the first operation. Creation is about 50 ms longer at p50, with a larger increase possible for uncached snapshots. The total time through the first operation remains the same. This change applies automatically, with no SDK upgrade or configuration changes required. See the Sandbox.create() documentation for details. Read more

Step 5 Preview now available on AI Gateway

October 8, 2026

Step 5 Preview, StepFun's flagship model, is now available on AI Gateway. It's built for agentic coding, professional knowledge work, and financial analysis, from building apps and debugging code to research and analytical reports. Step 5 Preview supports text and image input with a 1M-token context window, allowing you to work with large codebases, a stack of documents, or screenshots and charts in a single request. Use stepfun/step-5-preview as the model name: For other coding agents such as Claude Code, Codex, Cursor, and more, install the latest Vercel CLI and run setup: Then select stepfun/step-5-preview in the agent. See the coding agents guide for details. Try Step 5 Preview in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. You can view all language models available

Microfrontends Routing is now free for Firewall-mitigated traffic

October 7, 2026

Microfrontends Routing is now free for requests that the Web Application Firewall (WAF) denies, challenges, or rate limits. CDN Requests and Fast Data Transfer are already free for this traffic. The change applies automatically to every project using microfrontends. Learn how to set up WAF rules in the Firewall documentation. Read more

OpenAI Decisions API now available on AI Gateway

October 7, 2026

OpenAI's Decisions API is now available on AI Gateway through an OpenAI-compatible /v1/decisions endpoint. Use it with GPT-6 Luna Decisions or any other decision model on AI Gateway. Decision models answer typed questions about a shared input and return probabilities, choices, and scores instead of generated text, which suits routing, triage, guardrails, and rubric scoring. You can ask predicate (yes/no), choice, and score questions against the same input in one request. openai/gpt-6-luna-decisions is a separate model ID from openai/gpt-6-luna, which stays a language model for text generation. Requests to the decisions ID go to the Decisions API and are billed at its rates. Call it with the OpenAI SDK, the AI SDK, the existing HTTP API at /v1/evaluate, or the AI CLI. Upgrading decisions.create requires the OpenAI SDK 7.30.0 or later for JavaScript, or 3.26.0 or later for Python. In the AI SDK, experimental_decide requires ai 7.0.128 or later. OpenAI SDK and AI SDK Point the OpenAI SDK at AI Gateway and call decisions.create as you would against OpenAI. In the AI SDK, questions are keyed by name and the yes/no type is called boolean. Each key in questions becomes a key in answers: A

Claude Haiku 5.5 now available on AI Gateway

October 7, 2026

Claude Haiku 5.5 from Anthropic is now available on AI Gateway. It is built for high-volume, cost-sensitive tasks such as summaries, context compaction, database queries, and classification, and works as a subagent alongside larger Claude models on coding work. It is also the fastest Claude model at standard speed, which suits live customer support and browser use. Test it as a direct upgrade wherever you use Claude Haiku 4.5 today. Haiku 5.5 is the first Haiku model with effort levels. It uses adaptive thinking and supports low, medium, high, xhigh, and max effort, which set how much the model thinks and how many tokens it uses. You can turn thinking off at low, medium, and high. At xhigh and max, thinking must stay on. Haiku 5.5 supports Zero Data Retention on AI Gateway. It ships with biology and cybersecurity safeguards, so it may decline some requests in those areas. Call it through the AI SDK, the Chat Completions, Responses, and Anthropic Messages APIs, or a coding agent connected to AI Gateway. Set effort with reasoning in the AI SDK or reasoning_effort in Chat Completions. Use anthropic/claude-haiku-5.5 as the model name: To use it in Claude Code, Codex, Cursor, and more,

Glyph Cluster is now available in stealth for free on AI Gateway

October 7, 2026

Glyph Cluster is now available on AI Gateway as a stealth model for teams on Pro and Enterprise plans with purchased AI Gateway credits. It is free to use during the stealth period. Glyph Cluster is a reasoning model for coding and long-context analysis, and handles multi-step work such as planning, synthesis, quantitative reasoning, and comparing material across large inputs. For coding, Glyph Cluster can review code, explain failures, debug issues, propose changes, and reason through implementation options. It supports function calling, so agents can connect its reasoning to your own tools and systems, and it streams responses as they are generated. Glyph Cluster takes text input only, with no image or file input. Tool use is limited to function tools you define, and structured outputs are not supported. ZDR is not available for this model, and prompts and responses sent through it may be used for training and model improvement. Call it through the AI SDK or the OpenAI-compatible Chat Completions and Responses APIs, or select it in a coding agent connected to AI Gateway. Use stealth/glyph-cluster as the model name: To use it in Claude Code, Codex, Cursor, and more, install the la

Timestamp attributes are now supported in Vercel Flags

October 7, 2026

Vercel Flags now supports timestamp attributes for entities, which represent the users, teams, devices, or requests evaluated by a flag. Use timestamp attributes to show a campaign page only between two dates, run a Black Friday promotion, or target users who registered before a cutoff. Configure a timestamp attribute You can configure timestamp attributes in the dashboard or via the CLI. In the Vercel dashboard, open Flags, select Entities, and create or select an entity. Add an attribute and set its data type to timestamp. Or define the attribute with the CLI: Whether you configure the attribute in the dashboard or CLI, your application must provide its value when evaluating a flag. The entity and attribute names must match your configuration, and the value must be a Unix epoch timestamp in milliseconds. This Flags SDK example passes the current time as system.time. Date.now() returns the required value in milliseconds. Keeping it inside dedupe() ensures that every flag evaluated during the same request uses the same time: Add a targeting rule In the dashboard, open the flag and add a targeting rule using the timestamp attribute. The rule compares the value from your application

AI Gateway adds confidence-based decision fallbacks

October 6, 2026

AI Gateway can now escalate a decision request to a fallback model when the primary model's answer trips a confidence condition you set. Conditions can be combined, so an escalation can depend on more than one signal. The plain model names you already list in models keep catching outright errors, so existing fallbacks are unaffected. Confidence conditions cover Choice and Score questions. Boolean questions escalate on a probability range instead. Set the fallback by adding one conditional object to providerOptions.gateway.models. Requests without it keep their existing behavior. Leave out question and the condition checks every question of the matching type, so { "confidenceBelow": 0.6 } escalates when any Choice or Score answer falls below 0.6. Because a triggered fallback runs a second decision, it bills both stages. Decision fallbacks are in beta. Learn more in the decision fallbacks documentation. Read more

Nano Banana 2.1 now available on AI Gateway

October 6, 2026

Nano Banana 2.1 from Google is now available on AI Gateway. It improves image generation and editing over earlier Nano Banana models. This release targets product recontextualization, mask- and ink-based editing, and factuality. Product recontextualization places an existing product in a new scene. Mask- and ink-based edits change only the region marked by a mask or by strokes drawn on the image. Factuality is how accurately images depict real-world subjects. The model renders photorealistic skin tones, intricate materials, sharp lighting, and coherent backgrounds at the latency and cost of a Flash-tier model. Call it through the AI SDK or the OpenAI-compatible Chat Completions API, or the AI CLI. AI CLI is an open-source command-line tool from Vercel Labs for generating text, images, video, and audio from your terminal. It runs on AI Gateway, so any Gateway model ID works with -m. Use google/gemini-nano-banana-2.1 as the model name. Generated images are returned in result.files: Try Nano Banana 2.1 in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-

Mistral Large 4 now available on AI Gateway

October 6, 2026

Mistral Large 4 is now available on AI Gateway. The open-weight, natively multimodal model combines reasoning with text and image understanding. It is useful for cyber defense, manufacturing, and finance workflows. You can access the model through AI Gateway with one API key, without creating a separate Mistral account. Use Mistral Large 4 (mistral/mistral-large-4 ) with the AI SDK, OpenAI-compatible Chat Completions API, Responses API, or Anthropic Messages API. You can also select it in a coding agent connected to AI Gateway. For coding agents, run npx vercel ai-gateway setup and select mistral/mistral-large-4 in your agent's model settings. Try Mistral Large 4 in the model playground, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own provider key. Read more

OpenAI Ultrafast mode now available on AI Gateway

October 5, 2026

AI Gateway now supports OpenAI's Ultrafast service tier for GPT-6 Astra, providing faster output for interactive applications and rapid coding iterations. To use Ultrafast, request it for openai/gpt-6-astra through AI SDK, the Chat Completions API, or Responses API: For workflows with frequent tool calls, OpenAI recommends the Responses API over a persistent WebSocket connection to reduce overhead between turns. See the Ultrafast service-tier examples for persistent connections, including AI SDK over WebSocket. Ultrafast supports US and global processing. Requests pinned to unsupported regions, such as the EU, run at the standard (default) tier. Standard processing remains the default when no service tier is specified. Requests served at Ultrafast are billed at 6× the standard per-token rate, while requests that fall back to another tier are billed at the rate for the tier actually served. Check the GPT-6 Astra model page for current rates. Read more

FLUX 3 Image now available on AI Gateway

October 5, 2026

FLUX 3 Image from Black Forest Labs is now available on AI Gateway. Generate images from text or edit existing images using the same model, with one API key and no Black Forest Labs account required. FLUX 3 Image supports up to ten reference images for editing, combining subjects, or changing the composition. It supports five resolution tiers from 768 × 768 to 4K, with square, portrait, and landscape aspect ratios. Set AI_GATEWAY_API_KEY, then use bfl/flux-3-image as the model ID in AI SDK 7 or later. To generate an image, call generateImage with a text prompt: To edit an image, pass it in prompt.images with an instruction. This example edits the image generated above: You can also use FLUX 3 Image from the terminal using the same AI Gateway API key and model ID with AI CLI. Try FLUX 3 Image in the model playground, or see the image generation guide. Read more

Jev for Python engineers

October 2, 2026

It's simply impossible to not hear about Jev. Seemingly everyone is tinkering with it in some way, from using it to make trading decisions (what could possibly go wrong?) to generating UIs with it (maybe we're onto something here!). Jev is a new kind of AI model. You feed it data and ask it a set of multiple-choice questions, and it responds with its answers and how confident it is in each. Just like with any other model, Jev can and will make mistakes, but it will make them fast. ​ ​The bottom line is: Jev is built for making narrow decisions. Whether it's accurate enough for your use cases is something you'll need to test. The job of the Python team at Vercel is to make it easy for Python devs to tinker with Jev, among a few other things. So please welcome the latest release of the AI SDK for Python. Install it with uv add ai to use the new experimental evaluate() API, which talks directly to Jev's nervous system. Then create an AI Gateway key, set AI_GATEWAY_API_KEY, and try this complete example: Now, I'd like to explain what Jev is and how to make it work for you using a couple of examples. One is good and one is dumb. Ask Jev which is which. Let's have some fun! Jev ELI5 Jev

How Rogo ships agent-written code to production in 5 minutes on Vercel

October 2, 2026

Rogo on Vercel 73,000+ deployments in a single month Agent-written code shipped to production in 5 minutes 6 production AI agents automating work from churn analysis to deal desk Zero manual triage during production incidents with agent swarms on AI SDK Rogo builds Felix, an AI agent that generates decks, financial models, and reports for a majority of the world's largest financial institutions. The Rogo team is rearchitecting its internal applications on Vercel, where agent-written code ships in 5 minutes. When incidents hit production, swarms of agents on the AI SDK triage and remediate them without manual intervention. As a result, the team now pushes more than 73,000 deployments a month. Read more

Speed Insights deprecates First Input Delay on November 1st

October 1, 2026

Speed Insights is deprecating First Input Delay (FID) and will stop recording new measurements on November 1st. Speed Insights already uses Interaction to Next Paint (INP), the Core Web Vital that replaced FID, as the responsiveness metric in your Real Experience Score. The deprecation won’t require changes on your side, affect your score, or prevent you from viewing existing FID data. Learn more in the Speed Insights metrics documentation. Read more

Laya decision model now available on AI Gateway, free through October 31

October 1, 2026

Laya from Convai Innovations is free to use on AI Gateway for 1 month through October 31, 2026 via Boundless. Laya is an evaluation/decision model that evaluates shared context against yes/no, choice, or scoring questions and returns structured answers with probabilities. For example, use it to: Triage support requests: Detect a refund request and route it to the billing team. Route agent work: Choose which tool or workflow should handle a request. Check guardrails: Flag a request or response for review when it may violate a policy. Score against a rubric: Rate a ticket's urgency or an answer's quality on a defined scale. You can call Laya through the AI SDK evaluation API, the native HTTP API, or the TypeSafe-compatible API with convaiinnovations/laya or convaiinnovations/laya-free: Through October 31, 2026, both convaiinnovations/laya and convaiinnovations/laya-free are free when served by Boundless. The -free ID routes to Boundless and stops serving after the promotion; the standard ID begins billing then. Read the documentation on evaluation models on AI Gateway for more details, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and co

Microsoft AI models are now available on AI Gateway

October 1, 2026

Vercel and Microsoft AI (MAI) have partnered to make MAI models available on AI Gateway. AI Gateway is one of a limited number of platforms offering access to MAI models. MAI's focus on security and zero data retention (ZDR) align with Vercel's goal of giving AI Gateway users control over their data and transparency in how it is used for training. MAI models on AI Gateway MAI's latest audio models are now supported on AI Gateway. MAI-Voice-2.1 and MAI-Voice-2.1-Flash generate speech, while MAI-Transcribe-2 Streaming returns transcript updates as audio arrives. MAI-Voice-2.1 (microsoft/mai-voice-2.1) generates expressive speech in 23 languages and keeps a consistent speaker across longer passages. Use it for narration, audiobooks, podcasts, and lessons where delivery matters from start to finish. MAI-Voice-2.1-Flash (microsoft/mai-voice-2.1-flash) brings the same multilingual speech generation to lower-latency interactions. It is suited to voice agents, assistants, and spoken replies that need to arrive quickly. MAI-Transcribe-2 Streaming (microsoft/mai-transcribe-2-streaming) returns partial transcripts while audio is still arriving. Applications can show live captions or follow a

Voyage Rerank 3 and Rerank 3 Lite are now available on AI Gateway

September 30, 2026

Rerank 3 and Rerank 3 Lite from Voyage AI by MongoDB are now available on AI Gateway. The models reorder search results by relevance before passing documents to a language model. The series improves retrieval quality across domains, with the largest gains on long documents and code. Rerank 3 focuses on accuracy, while Rerank 3 Lite is optimized for latency and cost. Both handle up to 32K tokens per query-document pair and follow instructions in the query to guide ranking. To use Rerank 3, set the model to voyage/rerank-3: For Rerank 3 Lite, change the model to voyage/rerank-3-lite. Try Voyage Rerank 3 in the AI Gateway model playground, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own provider key. Read more

Vercel Agent now installs private packages from npm and custom registries

September 30, 2026

Vercel Agent can install private dependencies from npm and custom registries using credentials stored as shared environment variables on Vercel. npm, pnpm, and classic Yarn running in Agent sessions authenticate as they do in Vercel builds. To get started, add a shared environment variable for Development or Preview: Use NPM_TOKEN for private packages hosted on registry.npmjs.org. Use NPM_RC to configure custom or multiple registries. Vercel Agent reads only team-shared variables, not project-scoped ones. Credential values stay outside the sandbox, so the agent cannot read them. Learn more in the private dependencies docs. Read more

AI Gateway adds Browserbase Search and Fetch tools

September 30, 2026

Browserbase Search and Fetch are now available through AI Gateway. Give your models access to current information, find relevant pages, and retrieve their contents, alongside Browserbase's existing AI SDK and Vercel Marketplace integrations. AI Gateway lets you use Browserbase's tools across model providers with one API key. Add web search and page retrieval to any model that supports tool calling, and keep the same tools when you switch models. These helpers are available in AI SDK 7.0.116 and later. Update with pnpm add ai@latest and set your AI_GATEWAY_API_KEY. To add search and fetch, pass the helpers in tools . For pricing and configuration options, see the Browserbase Search and Fetch documentation. Try the Browserbase tools in the AI Gateway model playground, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own provider key. Read more

Edge Requests are now called CDN Requests

September 30, 2026

CDN Requests is now the name for the usage metric previously called Edge Requests. This is a naming change only: pricing, limits, and how usage is measured are all unchanged. The new name appears on the usage dashboard, the Observability Events chart, usage notifications, and invoice line items. It reflects that these are requests served by Vercel's CDN. Metric IDs and product aliases are unchanged, so saved queries keep working. In the billing charges API, ServiceName now returns the new name. Integrations that match charges on ServiceName should switch to the new SkuId field, which stays stable across renames. See the CDN pricing and usage documentation for details. Read more

Vercel Sandbox now supports Secure Compute

September 30, 2026

Vercel Sandbox now supports Secure Compute, connecting sandboxes to a team's dedicated network. Public-internet traffic exits through the network's static IPs, and sandboxes can reach private resources in an AWS VPC through VPC peering. You can attach a sandbox to your network in your code or via the CLI. In your code, pass an existing Secure Compute network ID when creating a sandbox: In the CLI, append --network-id and your network id value when you run the sandbox create command: Existing sandboxes can attach to or change networks with sandbox.update(). The change takes effect on the next session. A running session keeps its current network until it stops. Available to Enterprise teams with Secure Compute. Learn more in the Sandbox documentation. Read more

Ling 3.1 Flash is now available on AI Gateway

September 30, 2026

Ling 3.1 Flash from InclusionAI is now available on AI Gateway. The model is free to use through October 13, 2026. Ling 3.1 Flash is a hybrid reasoning language model with 560B total parameters and 25B active per token. It has a 262K-token context window on AI Gateway. The model is designed for coding, multi-step analysis, and agents that use tools, including workflows involving long documents, code, and extended task histories. To try Ling 3.1 Flash, use inclusionai/ling-3.1-flash or inclusionai/ling-3.1-flash-free as the model name: The standard model ID is free during the promotion and begins billing when it ends. The -free model ID stops serving instead of billing when the free promotion ends. To use Ling 3.1 Flash in a coding agent, follow the AI Gateway coding agents guide and select inclusionai/ling-3.1-flash as the model. Try Ling 3.1 Flash in the AI Gateway model playground, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own provider key. Read more

Vercel CDN no longer caches responses with Vary: Cookie

September 30, 2026

Vercel CDN no longer caches origin responses when Vary includes Cookie. Vary tells a cache which request headers may change the response. Because cookies often contain visitor-specific values, varying by Cookie can create many cache entries that are rarely reused. The response is still served normally but isn't stored for future requests. You can identify these responses by checking the x-vercel-cache header and Runtime Logs. The header is set to MISS, and Logs show vary_key_denied:cookie as the cache reason. If Logs show vary_key_denied:cookie for a route you expect to be cached, check the Vary header returned by your origin: If the response is the same regardless of cookies, remove Cookie from Vary. Vercel CDN can then cache it if it meets the other caching requirements. If the response depends on cookies, keep Cookie in Vary and use Cache-Control: private to prevent personalized responses from being stored in the shared CDN cache. Caching behavior for other supported Vary headers is unchanged. Learn more about high-cardinality headers and the vary_key_denied cache reason. Read more

Vercel Connect now accepts service submissions

September 29, 2026

Vercel Connect now lets you submit your own service to the connectors directory, so you no longer need to wait for Vercel to build a connector for it. Submitting a service takes four steps: Describe your service. Add one or more connection methods, using OAuth or an API key. Verify each OAuth method by requesting a real token through a test connector. Send the submission for review. Vercel reviews each submission before publishing it to the directory. Get started from Submit a Service in the dashboard, or read what your service needs to support in the documentation. Read more

Search trace spans from the Vercel CLI

September 29, 2026

You and your coding agents can now search a project’s trace spans from the terminal with vercel traces search. Each span represents one step in a request, helping you investigate errors, latency, and security-related behavior without opening the dashboard. By default, vercel traces search returns up to 100 spans from the last hour, newest first, so you can search without first finding a request or trace ID. Filter results by environment, service, span name, status, deployment ID, request ID, trace ID, or span ID, or narrow the time range with --since and --until: For filters without a dedicated option, use --query with Vercel's supported subset of Kibana Query Langugage (KQL). It supports fields such as duration, along with span and resource attributes: Add --json to output one JSON object per span, giving scripts and agents structured data to analyze. Update to the latest Vercel CLI with vercel upgrade, then run vercel traces search --help to see all available options. Learn more in the vercel traces documentation. Read more

GPT-6.1 Sol now available on AI Gateway

September 29, 2026

GPT-6.1 Sol from OpenAI is now available on AI Gateway. It improves on GPT-6 Sol for coding, computer use, and professional work, including reading complex documents and carrying out multi-step workflows. The model is suited to agents that need to debug code, work through business tasks, or extract answers from PDFs with tables and charts. It also improves factual accuracy on difficult questions. Its standard input and output pricing is lower than GPT-6 Astra's, while cheaper cached input makes it useful for requests that reuse a long shared context. Use openai/gpt-6.1-sol as the model name with the AI SDK, OpenAI-compatible Chat Completions API, or Responses API. You can also select it in coding agents connected to AI Gateway: To use the model in coding agents like Codex, Cursor, and more, install the latest Vercel CLI and run setup: Then select openai/gpt-6.1-sol in the agent. See the coding agents guide for details. Try GPT-6.1 Sol in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data

Search domains without authentication

September 28, 2026

You can now discover domain names and check availability and pricing without signing in to Vercel or providing an access token. This works through the Vercel CLI and Domains Registrar API. With the Vercel CLI, search by keyword or domain fragment: With the Domains Registrar API, check up to 200 exact domain names in one request. The response shows whether each can be registered and, when available, its registration and renewal prices: Authentication is still required to buy or manage domains. Try the public domain search, or learn how to claim a free domain with Pro. Read more

Claude Sonnet 5.5 now available on AI Gateway

September 28, 2026

Claude Sonnet 5.5 from Anthropic is now available on AI Gateway as anthropic/claude-sonnet-5.5. Sonnet 5.5 improves on Claude Sonnet 5 for well-scoped everyday work, from building features and fixing bugs to creating documents, slides, and spreadsheets. On coding tasks, Sonnet 5.5 carries changes through multiple steps and checks its work before reporting completion. It produces more polished documents, slides, and spreadsheets with fewer follow-up edits, and writes clearer explanations and reports. It also improves at using desktop apps and reading charts and technical drawings. Sonnet 5.5 supports Zero Data Retention on AI Gateway. Use Sonnet 5.5 with the AI SDK, OpenAI-compatible Chat Completions API, Responses API, or Anthropic Messages API. You can also select it in a coding agent connected to AI Gateway. For coding agents, run npx vercel ai-gateway setup and select anthropic/claude-sonnet-5.5 in your agent's model settings. Try Sonnet 5.5 in the AI Gateway model playground, or explore the model catalog. AI Gateway provides one API for calling models, tracking usage and cost, and configuring routing, retries, and failover. You can use an AI Gateway API key or bring your own pr

Ember-1 from Fireworks now available on AI Gateway

September 27, 2026

Ember-1 from Fireworks is now available on AI Gateway. Ember-1 is a research preview reasoning model built on Kimi K3 for coding and agentic workflows. Fireworks reports approximately 40% fewer generated tokens than Kimi K3 at comparable quality across its evaluations. For coding agents that make repeated model calls, shorter reasoning traces can reduce output costs and the amount of context carried into later steps. Ember-1 supports a 1M-token context window, text and image input, tool calling, and implicit prompt caching. The Fireworks endpoint supports Zero Data Retention and No Prompt Training. This model is available as a research preview with an initial two-week window. To get started, use fireworks/ember-1 for the model name across APIs and coding agents: To use Ember-1 in your coding agent, run the setup command with the Vercel CLI: The command detects installed agents, provisions or reuses an AI Gateway API key, and configures their connection to the gateway. Then select fireworks/ember-1 in your agent's model configuration. AI Gateway provides a unified API for calling models, with built-in usage and cost tracking, budgets for API keys, and routing rules. Try Ember-1 in t

Vercel Marketplace database browser now supports Redis

September 25, 2026

You can now run Redis commands, browse keys, and inspect values directly in the Vercel dashboard using the Marketplace database browser. Both the Redis and Upstash for Redis integrations in Vercel Marketplace are supported, so you no longer need redis-cli or a separate database UI. From your project, open Infrastructure > Storage and select a Redis database. The database page now includes two new tabs: CLI: Run Redis commands and inspect the results. Run multiple commands by separating them with newlines or semicolons, with support for MULTI/EXEC transactions. Browser: Browse keys in a flat list or grouped by prefix, and filter them by type or pattern. Select a key to inspect its value and metadata, including its type, TTL, encoding, and size. Supported value types include string, JSON, hash, list, set, and sorted set. These tabs are currently available only to team members with the Owner role. Learn more in the Marketplace storage documentation. Read more

Pixel Canary is now available in stealth for free on AI Gateway

September 25, 2026

Pixel Canary is now available on AI Gateway as stealth/pixel-canary. It's free for a limited time while in stealth. Pixel Canary is strong at coding, including building applications and refactoring existing code. It is well suited to frontend development and mobile app design, from responsive layouts and app screens to navigation and interactive components. On Next.js evals, Pixel Canary ties GPT 6 Astra (high) at a 90.3% baseline success rate, passing 28 of 31 tasks. With Next.js documentation supplied through AGENTS.md, it passes 30 of 31 tasks (96.8%), tying the leaderboard's top score in that setting. It passes tasks covering App Router migrations, data fetching, image and font optimization, caching, and view transitions, demonstrating its ability to implement and update the code behind web interfaces. Scores use pass@4: a task passes if any of up to four attempts succeeds. ZDR is not available for this model, and prompts and responses sent through it may be used for training and model improvement. To get started, set model to stealth/pixel-canary: To use Pixel Canary in your coding agent, run the setup command with the Vercel CLI: The command detects installed agents, provisio

Push images to Vercel Container Registry from GitHub Actions

September 25, 2026

You can now push container images from GitHub Actions to Vercel Container Registry (VCR) without storing long-lived registry credentials. The new vercel/vcr-action/login action authenticates your workflow using GitHub OIDC. It exchanges the workflow’s OIDC token for a short-lived Vercel access token, then uses that token to log in to vcr.vercel.com. When the job ends, the action logs out and revokes the Vercel token. To start: Create an OIDC policy on your Vercel team that matches the GitHub repository and workflow, and grants read-write access to VCR. Store your Vercel team ID as a GitHub repo variable, for example VERCEL_TEAM_ID, along with the team slug, project slug, and repo name used in the image tag. Give the workflow or job id-token: write permission. Then add the login step before you build and push and the image: The action authenticates Docker by default. Pass engines to use Podman or Buildah instead. After VCR finishes preparing a linux/amd64 image, you can use it as a custom Vercel Sandbox image. Within the same project, reference it as :. Learn more in the VCR documentation and the GitHub Actions guide. Read more

Vercel Sandbox now supports memory observability

September 25, 2026

Vercel Sandbox observability now includes memory usage data. You can access sandbox memory usage data in the dashboard and through the CLI via the vercel metrics command. Sandbox observability memory in the dashboard The Memory Usage card reports average, P75, and P95 memory across your sandboxes, alongside the existing CPU usage and data transfer metrics on the sandbox overview page and the project and team-level pages. On the sandbox detail page, charts are designed to show how close a sandbox is running to its memory ceiling by: Automatically scaling the y-axis to sandbox's memory limit Adding a dashed reference line at 85% of that limit Sandbox memory is also available in the Observability query builder as the memoryUsedBytes measure on the sandbox usage event. You can build custom queries against it and create alerts from the chart. Sandbox memory observability in the CLI Sandbox observability data is also available through the Vercel CLI via the vercel metrics command under vercel.sandbox.memory_used_bytes. Run the command in a linked project for project-specific data, or append --all to see aggregate data across all projects in your team. The CLI returns: The metric The repo

State of agent skills

September 25, 2026

In seven months, the skills.sh registry grew to one million agent skills and recorded nearly 280 million installs. A skill gives an AI agent reusable instructions for a particular job. Agents are capable but generic; they can do many jobs, but they don't know how a specific person, team, or company does them. A skill supplies that missing context and can be as simple as a file written in ordinary language. Anthropic introduced Agent Skills in October 2025, and Vercel launched the skills.sh registry three months later. Using aggregate data from the registry, this report follows a new market through its first million skills: what people teach agents, what gets installed, and where the value is heading. Skills turn expertise into software To write a skill, some of the judgment behind a job has to be made explicit, like the steps to follow, what good looks like, and how to evaluate the result. The agent and its tools already supply much of the underlying capability, so the skill only has to provide the judgment and instructions specific to a job. This makes skills faster and easier to create than conventional software. skills.sh reached one million skills in seven months. GitHub took 2

How Klaviyo shipped 356 internal apps in two weeks on Vercel

September 24, 2026

Klaviyo on Vercel Built a platform for shipping internal apps 512 builders and 356 live apps in two weeks Every app ships SSO-gated and private by default Builders go from idea to live app in 3 minutes Klaviyo is a B2C CRM for more than 200,000 brands. It brings marketing, service, agents, and data together in one platform so brands can deliver customer experiences at scale. Klaviyo built a platform on Vercel for shipping internal apps, then opened it to every employee through a citizen developer program. In the first two weeks of the program, 512 of Klaviyo’s more than 2,000 employees deployed 356 apps on Vercel. Of those, 196 were full-stack apps with their own databases. The teams that are now shipping software include legal, marketing, and HR. "The next big thing at Klaviyo could be in a marketer's laptop," says Sr. Lead AI Architect Mohamed Ali. Guardrails instead of walls Opening deployment to a couple thousand new builders made security the first conversation. The obvious way to keep them safe is to broadly restrict what they can do. Klaviyo went the other way. Employees aren't just building microsites; they are shipping full-stack apps on top of Klaviyo's own databases, and

The Vercel Bug Bounty Program is now publicly available

September 24, 2026

In 2022, we launched a private bug bounty program through HackerOne. Over the last several years, we've worked with HackerOne's VIP program to refine our process, targets, and scope, onboarding thousands of their best researchers and hardening security across our platform. We've translated those learnings into several public bounty programs: The news: [ million React2Shell hacker challenge The news: [ million Vercel Sandbox hacker challenge Our OSS bug bounty program Today, we are combining our private and OSS bounty programs into a single, public Vercel bug bounty program. Why now AI has fundamentally changed the nature of bug bounty programs, exponentially increasing the number of reports, both valid and invalid. Some companies are responding by moving to private programs. At Vercel, we’ve found that public reports are still surfacing real, valuable findings. AI enables a much wider range of researchers to discover vulnerabilities, and we believe in evaluating reports on their own merit, so making our entire program public was the logical choice. In preparation for the public launch, our security engineering team has streamlined our process and built tooling to filter out noise and expedit

Vercel Connect now supports TanStack AI

September 24, 2026

Agents built with TanStack AI can now call OAuth-protected MCP servers through Vercel Connect, with no credentials for you to store or rotate. The new @vercel/connect/tanstack-ai subpath exports connectMCPTransport, which takes a TanStack transport config and attaches a Connect-backed auth provider. The provider is called before every MCP request, so the token is always fresh. If the user has not granted access, createMCPClient fails with a consent challenge before the model runs. Catch it with getConsentChallenge and redirect to Connect's consent URL. Otherwise, a consent error raised within a tool call would reach the model as an error string rather than the user as a redirect. Read the TanStack AI integration docs to get started, or see the full list of supported frameworks and adapters for Vercel Connect. Read more

Gemini 3.8 text-to-speech models now available on AI Gateway

September 23, 2026

Gemini 3.8 Flash-Lite TTS and Gemini 3.8 Flash TTS from Google are now available on AI Gateway. Both models take text and generate speech in more than 100 languages. They support long-form narration, control over delivery, and two-speaker dialogue. google/gemini-3.8-flash-lite-tts is suited to high-volume speech generation, with controls for tone, pacing, and line-by-line delivery. google/gemini-3.8-flash-tts adds voice and character design through natural-language prompts, including acting cues, accents, and conversational reactions. Generate and listen to speech in the Flash-Lite TTS playground or the Flash TTS playground. For setup instructions and more code examples, see the speech quickstart and text-to-speech guide. AI Gateway provides one API for speech generation alongside your other models, with usage and cost tracking for each request. You can configure routing rules and bring your own provider key. Read more

Unlimited Vercel Blob stores on every plan

September 23, 2026

You can now create as many Vercel Blob stores as you need. The previous limits of 100 stores on Hobby, 500 on Pro, and 1,000 on Enterprise no longer apply. Blob store creation is now billed alongside other Blob Advanced Operations, including put(), copy(), and list() calls. On Pro that's $5.00 per million. On Hobby it counts toward the 2,000 free operations you get each month. Deleting a store is free. Create a new store whenever you want a hard boundary instead of a pathname convention: Separate production, staging, and preview data, and hand each environment its own credential. Create a store per customer in a multi-tenant app, so you can export or delete one tenant's data in a single call. Spin up a store for a preview branch or a migration, then delete it when you're done. Storage, operations, and data transfer are still billed on what you use, so splitting the same data across more stores costs the same. Store creation shows up under Blob Advanced Operations on your usage page and in the Observability dashboard. Learn more in the Vercel Blob Pricing documentation. Read more

GPT-6 Sol and Luna now available on AI Gateway

September 22, 2026

GPT-6 Sol and GPT-6 Luna from OpenAI are now available on AI Gateway. Both models bring GPT-6 improvements in professional work, coding, computer use, factuality, and communication at a lower price than GPT-6 Astra. GPT-6 Sol (openai/gpt-6-sol) is suited to complex professional workflows and sustained coding tasks where quality and room to iterate both matter. GPT-6 Luna (openai/gpt-6-luna) is the lower-cost option for high-volume agentic workflows, coding, and everyday tasks. Both Sol and Luna communicate more directly than their GPT-5.6 counterparts, with less jargon and fewer low-value details. They also improve factual reliability and are less likely to make misleading claims about work completed during coding tasks. Use openai/gpt-6-sol and openai/gpt-6-luna across the AI SDK, OpenAI-compatible Chat Completions API, OpenAI-compatible Responses API, and coding agents connected to AI Gateway. Install the latest Vercel CLI and connect your supported coding agents to AI Gateway: Then select openai/gpt-6-sol for longer or more demanding coding work, or openai/gpt-6-luna when throughput is the priority. AI Gateway provides a unified API for calling models, tracking usage, and config

Drives for Vercel Sandbox are now in public beta

September 22, 2026

Drives for Vercel Sandbox are now available in public beta on Hobby, Pro, and Enterprise. A Drive is persistent storage that you mount as a directory in a Vercel Sandbox. It isn’t tied to a single sandbox, so you can reuse the same Drive across runs and different sandbox instances. Use Drives to preserve an agent's workspace or on-disk memory, or to reuse datasets, models, and dependency trees. Create and mount a Drive Create or retrieve a Drive and mount it at a path when starting a sandbox. Read and write files through the sandbox filesystem at that path. Anything stored under /data remains on the Drive after the sandbox stops. Share Drive data across sandboxes A Drive supports one read-write mount at a time. After the Drive has been written to, multiple sandboxes can read from it concurrently by mounting point-in-time, read-only snapshots. Each snapshot reflects the Drive at the moment it's mounted. Later writes aren’t included; mount a new snapshot to access them. Limits and pricing Each sandbox can mount up to four Drives at separate paths. Drives default to a maximum size of 1 TiB (1 GiB on Hobby) and can be configured up to 16 TiB, with higher limits available by request. Dr

Claude Opus 5.5 now available on AI Gateway

September 22, 2026

Claude Opus 5.5 from Anthropic is now available on AI Gateway. It is a step-change improvement over Opus 5, with its biggest gains in agentic coding, long-running agent tasks, and knowledge work. Anthropic cites that Opus 5.5 performs at the level of Fable 5.1, but ~30% faster and ~40% cheaper than Opus 5 per task. Opus 5.5 is also a better collaborator over long runs. It reports back in plain language on what it did, what it found, and what it needs next, making it easier to supervise work that spans many steps or takes place over a longer period. Opus 5.5 includes two API changes that can turn previously valid requests into HTTP 400 errors: Thinking is always adaptive. Requests that disable thinking or set a fixed thinking budget are rejected. The model decides how much to think for each request. Use effort and prompting to steer its thinking behavior. Forced tool use is retired. Requests cannot require a tool call or force a specific tool. Prompt the model toward the tool, then catch and retry misses in your harness. If you previously forced a tool call to return JSON, use structured outputs instead. Use anthropic/claude-opus-5.5 across the AI SDK, OpenAI-compatible Chat Complet

Grok 4.7 now available on AI Gateway, fx, and eve

September 21, 2026

Grok 4.7 from SpaceXAI is now available on AI Gateway as spacexai/grok-4.7. Requests using this model ID are automatically 40% off through September 27. Grok 4.7 is designed for complex, long-running coding, document, and presentation tasks. It has a 500K-token context window and supports low, medium, high, and xhigh reasoning levels, giving you control over response speed and reasoning depth. To use Grok 4.7: To create an agent with eve, pass the same model ID and choose a reasoning level: Try Grok 4.7 in the model playground, see the coding agents guide for agent-specific setup instructions, or check out more discounted models on AI Gateway. Read more

MiMo V2.6 models now available on AI Gateway

September 21, 2026

MiMo V2.6 Pro, MiMo V2.6 Flash, and MiMo V2.6 Pro UltraSpeed from Xiaomi are now available on AI Gateway. MiMo V2.6 combines coding, reasoning, and tool use with native text, image, audio, and video understanding. Its 1M token context supports long repositories, tool traces, and multi-session agent work, with structured outputs and up to 128K output tokens. Choose a model based on the workload: xiaomi/mimo-v2.6-pro is the larger sparse mixture-of-experts checkpoint, with 1.02T total parameters and 42B activated per token, for complex software engineering and long-running agent work. xiaomi/mimo-v2.6-flash uses 309B total parameters and activates 15B per token, making it the more efficient option for multimodal automation and everyday agent workflows. xiaomi/mimo-v2.6-pro-ultraspeed serves Pro at up to 20 times its output speed for interactive and latency-sensitive workflows, with the same capabilities. To use MiMo V2.6 in a coding agent, install the latest Vercel CLI and run: Then select xiaomi/mimo-v2.6-pro, xiaomi/mimo-v2.6-flash, or xiaomi/mimo-v2.6-pro-ultraspeed in your agent. Try MiMo V2.6 Pro, Flash, or Pro UltraSpeed in the model playground. AI Gateway provides a unified AP

AI Gateway now supports TypeSafe clients and an HTTP API for Jev

September 21, 2026

You can now call Jev from TypeSafe AI through AI Gateway using an existing TypeSafe client or the HTTP API, in addition to the AI SDK. TypeSafe client: Point an existing TypeSafe client at AI Gateway without changing its evaluation calls. HTTP API: Call Jev directly from any language or framework. AI SDK: Call Jev from a TypeScript application using AI SDK. Jev is a probabilistic decision model for software. State goes in, and typed answers come out with probabilities attached, so there's no generated text to parse. Requests are billed through AI Gateway on all three paths, so they appear alongside your other model calls in usage and observability. Migrate an existing TypeSafe client Change the base URL and API key. Your systemOne calls, noul questions, and response shapes stay exactly as they are. Start a new integration New integrations name the model as typesafe-ai/jev and ask one of three question types: boolean returns a probability from 0 to 1, choice picks one option from a set you name, and score rates against a scale you define. This example asks whether an agent should keep working after fixing a bug and passing its tests. With the HTTP API, POST to /v1/evaluate: With the

Vercel Connect now supports Microsoft Teams

September 21, 2026

Vercel Connect now includes a managed connector for Microsoft Teams. Creating one gives your organization a Teams bot that your apps and agents run. People can @mention it in channels or message it directly, and your code receives the message and replies as the bot. As a Vercel Managed Connector, Vercel registers the Entra app and Azure Bot resource in your tenant, so there's no client secret to store. A tenant administrator with an Azure subscription completes setup once. Incoming Teams activities are verified and forwarded to your project as Connect trigger. Create a connector from the dashboard or Vercel CLI: Once the connector is set up, your code requests a token only when it needs one. Use it with the @vercel/connect SDK, eve channel or Chat SDK adapter: Each token is scoped to what you request and refreshed automatically, so there's nothing to rotate by hand. Connectors only work in the environments you attach them to, and you can revoke access at any time with vc connect revoke-tokens. Read the Vercel Connect documentation, view details about the Teams connector in the Vercel Connect catalog, or create a Teams connector to get started. Read more

Deployments now show billable duration and CPU minutes

September 21, 2026

Deployments now show their billable duration and CPU minutes in the dashboard, vc inspect, and the REST API. Use it to understand how each build contributes to your usage. Billable duration is the build and post-build time combined, rounded up to the next whole minute. CPU minutes are that figure multiplied by the machine's vCPUs. To see the same breakdown from the CLI, update to version 59.23.1 or later with npm i -g vercel@latest, then run vc inspect: Learn more about build machines and CPU minutes in the docs. Read more

Grok 4.7 now available and 40% off on AI Gateway, fx, and eve

September 21, 2026

Grok 4.7 from SpaceXAI is now available on AI Gateway and 40% off through September 27. The discount applies automatically when you call spacexai/grok-4.7. Grok 4.7 has a 500K token context window and supports low, medium, high, and xhigh reasoning levels, giving you control over the tradeoff between latency and depth. Use spacexai/grok-4.7 everywhere you call the model. The same ID works with the AI SDK, the OpenAI-compatible Chat Completions API, and coding agents connected to AI Gateway: To use Grok 4.7 in a coding agent, install the latest Vercel CLI and run setup: Then select spacexai/grok-4.7 in fx, Cursor, Codex, Amp, OpenCode, or another supported agent. See the coding agents guide for agent-specific instructions. To create a new eve agent with Grok 4.7 and xhigh reasoning, pass the same model ID to the initializer: Grok 4.7 supports Zero Data Retention and disallow prompt training. Requests appear in AI Gateway logs, custom reporting, and budgets alongside your other models. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in cus

Jev is the fastest-adopted model in AI Gateway history

September 18, 2026

Within 24 hours of launching on AI Gateway, Jev from TypeSafe AI reached more than twice as many paid teams as any previous model launch, making it the fastest-adopted model in gateway history. Jev passed every other comparison model in its first twelve hours and continued to widen its lead for the rest of the day. By hour 24, it was being used by nearly 13% of paid teams. That’s 2x as many as the GPT-5.6 family and over 6x as many as Fable 5.1. Jev’s launch shows how quickly a specialized model can find a place in production. Its first-day adoption was unmatched among recent launches; the next test is whether that early adoption lasts. About Jev Jev was introduced on September 15 as a probabilistic decision model built for making structured decisions inside software. An application sends it context and a set of questions. Jev evaluates those questions in parallel and returns typed choices, scores, or true-or-false answers, along with probabilities. Unlike the text produced by a general-purpose language model, Jev’s answers come in a format the code can use directly. Developers can use it to: choose an agent’s next tool or subagent decide whether a workflow should continue, retry,

Reproducing, disclosing, and fixing the libheif vulnerability with Hacktron and the maintainers

September 18, 2026

In August 2026, Hacktron reported what looked like a remote code execution (RCE) vulnerability in Next.js image optimization. Their investigation found that the vulnerable code was not in Next.js itself, but upstream in libheif, an AVIF image decoder used by Next.js, ImageMagick, WordPress, sharp, and much of the web. Shortly after Hacktron notified us, we worked with them to reproduce the RCE against a current Next.js build and disclose it to the maintainers of sharp, libvips, and libheif. We then deployed a platform-wide mitigation on Vercel and started working with the maintainers on a fix. The dependency chain Next.js image optimization lets applications resize and optimize images through the component (next/image). For AVIF images, the image-processing dependency chain is as follows: invokes /_next/image, /_next/image calls sharp sharp calls libvips libvips uses libheif to decode the image That meant the vulnerable code was not in Next.js, but it was still reachable through Next.js image optimization. A malicious AVIF image sent to the image optimization endpoint would invoke libheif through sharp and libvips. As such, one obvious mitigation was to disable AVIF optimization in

Spend Management expands to Enterprise Flexible Commitment plans

September 18, 2026

Enterprise teams on Flexible Commitment plans can now use Spend Management, already available on Pro, at no additional cost. You can set a budget at any time in Spend Management settings. Set a budget per billing cycle, and when your team's metered usage approaches or crosses it, Spend Management can send email notifications, trigger a webhook, or pause the production deployments of all projects. The amount governs the metered usage that draws down from your prepaid balance. Setting an amount does not stop usage on its own; pausing is opt-in, and it does not stop AI Gateway or v0 usage. Learn more in the Spend Management docs. Read more

WebMCP support now available in mcp-handler

September 18, 2026

mcp-handler now has experimental support for WebMCP, the proposed web standard for exposing tools to in-browser agents. Add a single script tag to your site, and your existing MCP tools become available there too. Opt tools in by adding them to the experimental_webMcp object: Then load the script from your MCP endpoint with the ?webmcp-script parameter: The script registers those tools with the page and proxies each call back to your MCP server as the signed-in user, so authenticated tools work without a browser-side OAuth flow. Upgrade to mcp-handler@2.2.0 and read the documentation to get started. Read more

v0 now reads npm credentials from shared environment variables

September 18, 2026

v0 now installs private packages from npm and custom registries using credentials stored as shared environment variables on Vercel. This makes it easier for teams to build with their existing design systems, component libraries, and internal packages directly in v0. To get started, add one of the following as a shared environment variable on Vercel, scoped to Development and/or Preview: Use NPM_TOKEN for private packages hosted on registry.npmjs.org. Use NPM_RC to configure custom or multiple registries. NPM_RC supports scoped registries and references to other environment variables. For example, configure an organization scope such as @acme for GitHub Packages, or direct package requests through a private JFrog Artifactory registry. Credentials can be marked sensitive, and v0 never exposes them to the model or writes them to the sandbox filesystem. You can view the integration status from Settings → Integrations in v0. Learn more in the private dependencies docs. Read more

GLM 5.3 FlashX now available on AI Gateway

September 18, 2026

GLM 5.3 FlashX is now available on AI Gateway. GLM 5.3 FlashX is a high-speed serving option for Z.ai's multimodal coding model, delivering inference at ~200 tokens per second for faster streamed responses. The higher serving speed is useful for coding agents, tool loops, and interactive applications where users wait on generated output. Use zai/glm-5.3-flashx across API formats and in coding agents: To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway setup to create a key and configure your supported agents. Select zai/glm-5.3-flashx inside the agent. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try GLM-5.3-FlashX in the model playground, or view all language models available on AI Gateway. Read more

GPT-Live 1 now available on AI Gateway

September 17, 2026

GPT-Live 1 from OpenAI is now available on AI Gateway. GPT-Live 1 is a full-duplex voice model and can listen and speak at the same time. Many voice models use turn detection to respond. Full duplex removes that boundary, so a user can pause, interrupt, or add detail while GPT-Live is speaking. GPT-Live 1 also supports client delegation. Client delegation lets you choose the background model independently from GPT-Live 1. A basic session runs only the voice model. When deeper work is needed, your application can handle delegation-created with any text model on AI Gateway while the conversation continues. Install AI SDK 7, @ai-sdk/openai 4.0.67 or later, and a WebSocket client: Simple example without delegation This starts openai/gpt-live-1 without calling another model. Wait for session-started before sending audio. Delegate work and keep talking You can call a text model for GPT Live to delegate to, then return the result on the commentary channel for it to speak. This example uses openai/gpt-5.6-sol, but you can substitute any text model available on AI Gateway: Your application controls delegated work and its permissions, confirmations, and cancellation. Delegated model requests

Open-weight models take 56% of token volume, Astra doubles Fable 5.1 spend

September 17, 2026

AI Gateway Production Index — September 2026 Every month, AI Gateway routes tens of trillions of tokens between production applications and AI labs. That traffic gives us a view of what AI usage actually looks like in today's enterprise, and we publish it here monthly. See the Production Index reports from June, July, and August. September 2026 summary The September index reports on AI Gateway data collected through August 2026. Open-weight models ran the majority of gateway tokens for the first time, up from 7% in December to 56% in August. The average token costs less than half what it did five months ago. Price per token fell 23.2% in August, the third straight monthly drop, and the median team paid 7.6% less. Fable 5, Anthropic's most capable model, lost two-thirds of its share of gateway spend in one month. Opus 5, at half the price, tripled its share. Anthropic kept 64% of all spend. Gemini 3 Flash has lost 95% of its share of gateway tokens since May, and more than three-quarters of the volume it lost went to models from other labs. Special report: OpenAI launches Astra on September 3 GPT-6 Astra took a third of OpenAI's spend within two days of launch and twice Fable 5.1's

Sub-second artifact deployments are now supported in Vercel CLI

September 17, 2026

You and your agents can now deploy static artifacts to Vercel in under one second through Vercel CLI. Run vercel deploy to share a prototype, publish an HTML report, or preview a page created by your coding agent. Vercel automatically detects eligible deployments, and valid artifacts will skip the build step, immediately returning a live URL. Instant deployments support directories with: Up to 10 HTML or Markdown files A total size of 5 MB or less Supported file extension include .html, .htm, and .md. Update to Vercel CLI version 59.16.0 or later to get started, and run vercel deploy --help for all available options. Read more

Turbo build machines can now be enabled per deployment

September 17, 2026

You can now opt into Turbo build machines on any individual deployment. This is useful when you need to increase resources temporarily without changing project settings. You can do this in three ways: Include #VERCEL_BUILD_MACHINE=TURBO in your Git commit message before pushing to GitHub Use vc deploy --turbo (Vercel CLI 59.20.0 or later) Set buildMachine to turbo when creating a deployment with the REST API Learn more in the managing builds documentation. Read more

Run Terminal-Bench and other Harbor evals on Vercel Sandbox

September 17, 2026

You can now run Harbor evals on Vercel Sandbox. Harbor is the open-source harness behind Terminal-Bench, whose registry includes many other benchmarks such as SWE-bench, tau3-bench and OSWorld. Pass --env vercel to harbor run and each trial executes in its own isolated Firecracker microVM, so you can parallelize far beyond what your local machine is capable of. A task's network policy is enforced at the sandbox firewall, outside the VM. Optional credential injection attaches secrets to matching outbound requests at that firewall, so they never enter the sandbox. Paired with AI Gateway, one AI_GATEWAY_API_KEY reaches hundreds of models from multiple providers, and benchmarking another model is the same command with a different --model: Swap --model to vercel_ai_gateway/openai/gpt-5.6-luna to run the same benchmark against an OpenAI model. Requires Harbor 0.22.0 or later. Follow the step-by-step guide for setup, configuration, and troubleshooting. Learn more in the Sandbox documentation. Read more

The skills CLI now supports Notion hosted skills

September 17, 2026

skills@1.7.0 adds Notion skills databases as an install source for agent skills . Notion skills are reusable agent skills written as Notion pages. Teams author, review, and update them in the workspace they already use, then install them into any agent the skills CLI supports. No Git repository required. To browse your Notion workspace's skills, run: The CLI lists the skill packs shared with you and installs every skill in the packs you select. To install a single skill, pass its Notion page URL: Both commands use the Notion CLI (ntn) to authenticate. To set it up: ntn login requires a Notion personal access token, so your workspace must allow them. Access follows Notion's page permissions. You only see skills shared with you, so controlling who can install a skill is the same as controlling who can view the page. This integration is built on Notion's new Agent Skills API, which exposes skills stored in Notion as standard Agent Skills folders. Because the format is standard, the same skills work in any agent that reads them. Get started by creating a Notion skill. Read more

Native Marketplace integrations now support custom environments

September 17, 2026

You can now connect native Marketplace resources to custom environments. Previously, resource connections could only target production, preview, and development environments. Choose custom environments when connecting a resource from the Vercel dashboard, Vercel CLI, or REST API. Vercel scopes the environment variables created by the connection to the selected environments. For example, from the CLI: Existing deployments do not change. Create a new deployment after connecting a resource or changing its environment scope. Custom environments are available on Pro and Enterprise plans. Read the docs for connecting from the dashboard, the CLI --environment flag, and the REST API envVarEnvironments field. Read more

TypeSafe AI's Jev now available on AI Gateway

September 16, 2026

Jev from TypeSafe AI is now available on AI Gateway. Jev is a probabilistic decision model for software: state goes in, typed Choice, Score, and Boolean answers come out. Regular language models generate text one token at a time, which the application then parses and validates. Jev evaluates all declared questions in parallel and returns typed answers plus probabilities directly. That removes unnecessary text generation and makes it straightforward to automate clear cases while routing uncertain ones to review. TypeSafe reports Jev was up to 193.6x faster and 444.6x cheaper than LLMs on its workflow evaluations. Example use cases include: Choosing the next tool or subagent in an agent loop Deciding whether to continue, retry, ask the user, or stop Scoring urgency or risk before an action Verifying model outputs and enforcing guardrails. AI SDK 7 exposes Jev through the experimental evaluate API. Choice selects an option, Score grades an ordered rubric, and Boolean estimates the probability of true. Install the current AI SDK (AI SDK 7.0.105 onwards supports the evaluate API): Each evaluation specifies: model: the evaluation model to call, state: the shared string, object, or array

Hobby projects now retain fewer deployments to free up storage

September 16, 2026

Hobby projects now retain fewer deployments past the 30-day retention window. Hobby teams get 10GB of Deployment Storage. Every deployment you keep uses some of it, and going over the limit can block you from deploying until you free some up. Deployment Retention for Hobby teams now deletes old deployments sooner, so dormant projects stop holding storage that your active projects need. Each Hobby project now keeps its 3 most recent production deployments, plus its 3 most recent deployments of any type, regardless of age. Preview deployments no longer get their own protection. Your current production deployment is still never deleted, and aliased and active-branch deployments are still protected. See the full list of exceptions in the docs. If your team is over the 10GB limit, deployments outside those exceptions are now deleted immediately instead of after 30 days. To stay under the limit, see how to optimize your Deployment Storage usage, or upgrade to Pro, where storage is billed at $0.10 per GB per month. Read more

Secure Compute and Static IP builds start 64% faster

September 16, 2026

Builds using Secure Compute or Static IPs now start 64% faster, with the average time from deployment creation to build start dropping from 6.7 seconds to 2.4 seconds. Previously, each build waited for a new build container to boot with its network configuration. These builds now use prewarmed build containers, with your network configuration attached when the build starts. The improvement is applied automatically to builds using Secure Compute or Static IPs, with no configuration changes required. Learn more about Secure Compute and Static IPs. Read more

Mem0 joins the Vercel Marketplace

September 16, 2026

Mem0 is now available as a native integration on the Vercel Marketplace, giving your AI agents and apps long-term memory. Mem0 remembers user preferences, facts, and context across sessions, so your app stops starting from scratch. Install from the Marketplace with integrated billing on your Vercel invoice, no separate account or key management. A scoped Mem0 project and API key are provisioned automatically and added to your project as environment variables, including MEM0_API_KEY. Mem0 offers a free plan with usage limits and a $20 per month plan for higher volume. To see it end to end, deploy the eve Memory Agent template, an eve agent with long-term memory powered by Mem0. Get started with Mem0 on the Vercel Marketplace or through the Vercel CLI, available to customers on all plans. Read more

Is Agentic now tailors its audit by site type

September 16, 2026

Is Agentic reports now let you view your checks through one of four site types: Docs & content, Business, App, or Commerce. For example, the Commerce view highlights payment and checkout standards like x402, UCP, and ACP, while the App view highlights API discovery, authentication, error handling, and SDK support. Your score stays the same in every view, and you can change views on an existing report without rescanning. Only the highlighted checks change, so you can focus on what matters for your kind of site while keeping scores comparable across sites. To set the default view, add a single meta tag to your page's : Set the value to match your site type: content for Docs & content business for Business app for App store for Commerce The tag only sets which view readers see first. It doesn't add points or change your score. If no type is declared, Is Agentic infers one and shows which type it chose. Run a check at is-agentic.com or read the documentation. Read more

How Delphi ships 100 times a day with its Python backend on Vercel

September 15, 2026

Delphi on Vercel 10 engineers with no dedicated infrastructure role Everyone ships code, including product and design 100+ production deploys a day behind feature flags Delphi builds digital minds. They capture what someone has written, recorded, and taught, so anyone can tap that expertise on demand. Delphi is “a destination where you can find and interact with interesting minds,” says founding engineer Spencer Schoeben. Nobody has built a digital mind network before, so Delphi figures out what to build by shipping features and watching how people use them. That only works if the product is fast to change. "If we are going to win, we need to have an excellent developer and agentic experience," Spencer says. "Otherwise, we could not adapt to how fast our user’s demands change." Delphi's frontend had run on Vercel since day one. "The frontend has always been easy," Spencer says, "but the backend was a nightmare." Six months ago, the team rebuilt its Python backend on Vercel as well. Today the whole team ships to production 100+ times a day. Building a Python backend that's easy to work on The backend slowed every change Delphi's backend previously ran on AWS, with ECS, Docker Deskto

Gemini 3.8 Live models now available on AI Gateway

September 15, 2026

Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking from Google are now available on AI Gateway. Both models support real-time spoken interactions for voice assistants, conversational experiences, and applications that respond through audio. google/gemini-3.8-live supports real-time audio, visual grounding, automatic switching across 97 languages, and background tool calls while the conversation continues. google/gemini-3.8-live-extended-thinking adds multi-step reasoning that runs in parallel with speech, allowing it to acknowledge requests and narrate progress without interrupting the conversation. Use either model through the AI SDK's realtime API. Install the Gateway provider and a WebSocket client: Mint a short-lived token, open the WebSocket, and use the model adapter to serialize and parse realtime events: See the realtime quickstart for more details on realtime events and WebSocket connections. Try Gemini 3.8 Live or Gemini 3.8 Live Extended Thinking in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. Read more

AI SDK harness layer now supports native subscription authentication

September 14, 2026

The AI SDK harness layer now supports authenticating harnesses through their native subscriptions, where the underlying harness supports them. The harness layer runs different coding agents through the same HarnessAgent interface, so you can switch agents without changing your application code. No code changes or new settings are required. The direct authentication mode uses explicit provider environment credentials when they are present, and otherwise a native subscription found on the host. The default auto mode does the same when no AI Gateway credentials are set. The ai-gateway mode never reads native subscriptions. Native subscription credentials stay on the host. Just like with API keys, credentials are resolved at the host boundary. OAuth access tokens are refreshed there as needed. Where the sandbox supports it, the harness receives placeholder credentials and the real token is injected into outbound requests on the host. Native subscriptions work across harness adapters wherever the harness itself supports subscription login. These include Claude Code, Cline, Codex, Cursor, fx, GitHub Copilot, Grok Build, OpenCode, and Pi. Read the harness documentation to learn more. Read

How Featured's users make 100K media pitches per month on Vercel

September 11, 2026

Featured on Vercel 3 engineers supporting 3 brands and 100,000+ users on Vercel Migrated 374 Sanity sites from AWS Elastic Beanstalk to Vercel AI SDK and AI Gateway power Featured's chat bot across 17 models Workflow SDK replaced custom long-running job infrastructure Featured is a co-pilot for public relations (PR) that subject matter experts and PR teams use to find media opportunities. Tell Featured's agents what you know, and it surfaces opportunities across journalist requests, podcasts, awards, and GEO, with no PR background required. Founder Brett Farmiloe knows from experience how hard and time consuming getting press is. He spent 10 years running Markitors, a digital marketing agency with 500 small business clients. Every client, from an eyelash extension supplier to an equipment financing company, had real expertise to share, but no way to get it in front of journalists. PR, as Farmiloe puts it, "has always been about who has access to what." He founded Featured to change the question from who has access to who has knowledge. Featured connects one of their users with a journalist or publisher every 6 seconds. Their agents deliver more than 100,000 media pitches per month,

How Tailscale built a customer-facing model router on AI Gateway

September 11, 2026

Tailscale on Vercel Hundreds of AI models shipped to customers in-product Model access granted and revoked by tailnet network identity Went from model routing prototype to paying customers in months Tailscale connects a company's laptops, servers, cloud instances, and personal devices into one private network called a tailnet. Remy Guercio, who leads product for Aperture by Tailscale, describes it simply: "It's basically like a VPC that can span any cloud, on-prem, your house, and your phone." Aperture takes that same idea and applies it to AI. Instead of giving every employee, agent, or tool a separate provider API key, Aperture lets companies control model access through the tailnet itself. Add someone to the network, and they can immediately use approved models. Remove them, and access disappears. Under the hood, Aperture is built on Vercel AI Gateway and Vercel Sandbox. AI Gateway gives Tailscale one API for hundreds of models. Sandbox gives agents a safe place to run. Together, they let Tailscale offer model access and agent execution inside a customer's private network, without their team building every piece of AI infrastructure from scratch. Model routing is harder than it

Control who can manage connectors in Vercel Connect

September 11, 2026

Pro and Enterprise teams can now control who can create and manage Vercel Connect connectors. Connectors let applications and agents access external services using credentials managed by your team. Owners can enable this restriction under Connector Permissions in Team Settings. Once enabled, only people with the Owner role or Connector Manager extended permission can create or manage connectors. Browse all connectors, or learn more in the Vercel Connect documentation. Read more

Vercel Sandbox now provides 64 GB of storage

September 11, 2026

Every Vercel Sandbox now includes 64 GB of storage, up from 32 GB. This includes sandboxes created from a Vercel Managed Image or custom image, as well as those configured with the deprecated runtime property. The additional space provides more room for large repositories, dependencies, build artifacts, and data workloads that spill to disk. Learn more about Sandbox images in the documentation. Read more

GitHub Copilot is now available in the AI SDK harness layer

September 10, 2026

The AI SDK harness layer now supports GitHub Copilot through the official @ai-sdk/harness-github-copilot adapter. The harness layer lets your application run different coding agents through the same HarnessAgent interface, so you can switch agents without changing your application code. Pass githubCopilot to HarnessAgent: Under the hood, the adapter uses @ai-sdk/harness-acp to connect GitHub Copilot to HarnessAgent through the Agent Client Protocol (ACP). Supported harnesses now include, in addition to GitHub Copilot, Claude Code, Cline, Codex, Cursor, Deep Agents, fx, Grok Build, OpenCode, and Pi, with more coming soon. Read the GitHub Copilot harness documentation to get started. Read more

Tako Search is free on AI Gateway through September 30

September 10, 2026

Tako Search is free exclusively on AI Gateway through September 30. It lets AI models search Tako's curated data and the live web, filter web results by domain or publication date, and use the results to answer questions with current information, citations, and visualizations. After September 30, searches are billed at standard rates. The same integration works with any model on AI Gateway, so you can switch models without changing your search setup. You also don't need a separate Tako account or API key. To use Tako Search with the AI SDK, add gateway.tools.takoSearch() to a generateText or streamText request. The model can then call it when it needs current information: Try Tako Search in the AI Gateway playground. See the web search documentation for configuration and search options. Read more

Build with OpenAI Agents API on Vercel

September 10, 2026

OpenAI Agents API provides an OpenAI-hosted agent harness for long-running, tool-using workflows. You can host the agent experience on Vercel and connect each session to Vercel Sandbox for code execution and file access. OpenAI runs the agent loop and maintains session state. Signed OpenAI webhooks are processed through Vercel Queues, while Vercel Sandbox provides an isolated, persistent execution environment for each agent session. With this integration, you get: An OpenAI-managed agent loop and session state Durable processing of sandbox lifecycle events An isolated execution environment for every agent session Persistent workspaces that remain available across follow-up instructions A scale-to-zero architecture without long-lived virtual machines Follow the step-by-step guide to build and deploy an agent, or explore the sample application. Read more

FastAPI frontends and static files served from the CDN

September 10, 2026

FastAPI frontends and static files, served with app.frontend() or StaticFiles, are now promoted to the Vercel CDN at build time. Requests for those paths are served directly from the CDN, without invoking your Vercel Function. FastAPI evaluates routes in declaration order. A route declared before a StaticFiles mount takes priority over any CDN file at that path. This precedence is preserved. Promoted source directories remain in the function bundle so the app can read from them at runtime. To exclude them and serve only from the CDN, set tool.vercel.fastapi.static.exclude = true in pyproject.toml. Frontends guarded by dependencies are kept on the function as the CDN cannot run dependency checks. Similarly, static files which sit behind middleware are also kept on the function. To override this and always promote to CDN, set tool.vercel.fastapi.static.cdn = true. To opt out of the CDN entirely, set it to false instead. For more information see the Vercel FastAPI documentation. Read more

How we cut CDN metadata lookup latency by 91%

September 10, 2026

Every request to Vercel passes through our CDN, which executes on average over 80 million routing instructions per second. Part of that work is looking up metadata to determine which paths exist and how to serve them. When that metadata isn’t cached, the CDN has to fetch it before it can serve the response. We used to fetch and cache metadata one path at a time, retrieving only what each lookup needed. That seemed efficient at the time, but large deployments can contain hundreds of thousands of paths, each with its own cache entry. Every new deployment introduced fresh metadata, making cache misses a recurring cost for large sites that deployed frequently. Fetching more metadata at once made these lookups faster. By grouping paths together, each fetch could populate the cache for many subsequent lookups. But fetching too much introduced its own costs. Through production experiments, we found a balance that cut P99 metadata lookup latency by 91% and made deployments faster along the way. How the CDN finds the right route The path in a request doesn’t always match the path of the content or function that serves it. Framework routing rules connect the two. A request for /blog/hello-wo

Tako Search is free on AI Gateway through September 30th

September 10, 2026

Tako Search is free exclusively on AI Gateway through September 30th. Tako Search is a model-agnostic search tool that works with any model on AI Gateway. You can switch models while keeping the same search integration, without a separate Tako account or API key. Tako Search combines its curated knowledge graph with the live web in one response, giving models source-grounded data cards, citations, and visualizations. You can search curated data, the live web, and narrow web results by domain and publication date. To use it with the AI SDK, add gateway.tools.takoSearch() to the tools object in a generateText or streamText request. The model can call Tako when it needs external information. AI Gateway runs the search and returns the results to the model to inform its answer. Try Tako Search in the AI Gateway playground. See the AI Gateway web-search documentation for setup, configuration, and search options. Read more

Vercel Sandbox is now available in all regions

September 10, 2026

Vercel Sandbox can now run in all 20 Vercel compute regions, up from four. Running sandboxes closer to the databases, storage, and other services they access reduces latency. Broader regional coverage also gives teams more control over where workloads execute, helping align them with data residency and regional processing requirements. iad1 remains the default region. Region selection is available on every plan. Pro and Enterprise teams can also configure failover regions. If Vercel can't create a sandbox in the primary region, it tries each configured failover region in order. Teams with data residency requirements can limit primary and failover regions to approved geographies. Set a default region for new sandboxes from your project's Settings > Sandboxes. With the SDK: With the CLI: A region passed at creation overrides the project default. Existing sandboxes remain in the regions where they were created. Active CPU and Provisioned Memory rates vary by region. See Sandbox pricing. If you use the Sandbox SDK or CLI, update it to the latest version before selecting one of the newly supported regions. Learn about region selection, failover, and CLI configuration in the Sandbox regi

DeepSeek V4.1 Flash now available on AI Gateway

September 9, 2026

DeepSeek V4.1 Flash is now available on AI Gateway with native image understanding. V4.1 Flash has vision support and accepts text and images in the same request, so you can ask questions about screenshots, read charts, and extract information from visual content. The model has a 1 million token context window and supports responses up to 384,000 tokens, along with reasoning, tool use, and prompt caching. Its new architecture processes input and generates output with separate components, reducing the active computation needed for each stage. Use deepseek/deepseek-v4.1-flash as the model name: To use it in Claude Code, Codex, Cursor, and more, install the latest Vercel CLI and run setup: Then select deepseek/deepseek-v4.1-flash in the agent. See the coding agents guide for details. Try DeepSeek V4.1 Flash in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not ch

You can now read and search changelogs from the CLI

September 9, 2026

You and your agents can now read and search the Vercel changelog feed from your terminal using vercel changelog. Coding agents can use this command to discover new Vercel products and features, find updates relevant to your project, and read full announcements to inform their recommendations. vercel changelog returns the latest five changelogs with full Markdown content. You can also set the number of results, search all changelogs using a keyword, and access JSON output for scripts and agents. Update the Vercel CLI to version 59.6.0 or later and run vercel changelog to get started. See vercel changelog --help for all available options. Read more

Password Protection is now available per project on Pro

September 9, 2026

Pro teams can now enable Password Protection for individual projects at $20 per project per month. Previously, it was available only through a $150-per-month team-level add-on that covered every project. When enabled, Password Protection requires visitors to enter a password you set before they can view the project’s deployments. Open Security in the sidebar, select Deployment Protection, and turn it on. Disable it to stop future charges for that project. Learn more about Deployment Protection usage and pricing in the documentation. Read more

Persistent memory for eve agents

September 9, 2026

eve agents can now retain context across sessions and use it in future conversations. Persistent memory is organized into slots. You can define named slots in files under agent/memory/. Each slot specifies a provider, which stores and retrieves the memory, and a scope, which determines who or what shares it. For example, you can keep separate memory for each authenticated user. Before each turn, eve retrieves relevant memory and adds it to the model's context. Depending on the provider, memory can be updated automatically after a turn, through tools the agent uses, or through both methods. To add eve's built-in file memory provider, run: The command creates a file memory slot scoped to each authenticated caller: For agents deployed on Vercel, file memory uses a private Vercel Blob store so saved memories persist across restarts and deployments. Other supported memory providers include Supermemory, Upstash AgentKit, and Kybernesis Arcana. You can also connect another storage system or memory service by building a custom provider. Read the memory documentation to choose a provider and get started. Read more

v0 adds one-click integrations for email, auth, search, and databases

September 9, 2026

We're working toward bringing parity across Vercel integration and v0, starting with Resend, Amazon OpenSearch, MongoDB Atlas, Algolia and Clerk. Prompt v0 with what you want to build, and when your prompt requires a provider, v0 renders a connect card in the chat. With Marketplace integrations in v0, you get: Connect as you build: Prompt v0 with what you want to create, and connect the required provider inline. Automatic setup: Once connected, v0 handles the required environment variables and configuration. Provider skills loaded automatically: When you connect a provider that publishes agent skills, v0 loads them and generates code that follows the provider's recommended patterns. Provider-specific capabilities: Resend, for example, adds transactional email with React Email components. To get started, open a v0 chat and prompt what you want to add to your stack. Read more

Password Protection is now priced per project on Pro

September 9, 2026

Pro teams can now enable Password Protection for $20 per project per month. It requires visitors to enter a password you set to view the project's deployments. Previously, Pro teams needed to purchase password protection for every project at the team level, but now it can be purchased individually. To enable Password Protection for a project, open Security in the sidebar, select Deployment Protection, and turn it on. Disable it to stop future charges for that project. Learn more about Deployment Protection usage and pricing in the documentation. Read more

Protect production deployments for free on every plan

September 9, 2026

Vercel Authentication can now protect every deployment URL for a project, including its production domains, at no additional cost on Hobby, Pro, and Enterprise. Previously, protecting all deployments with Vercel Authentication required the $150-per-month Advanced Deployment Protection add-on. When enabled, only visitors who sign in to Vercel and have access to the project can open protected deployments. Use it for internal tools, private dashboards, and sites that aren’t yet ready to launch. From your project, open Security in the sidebar, select Deployment Protection, then choose All Deployments from the dropdown. You can also set this configuration as your team default, so Vercel Authentication protects all deployments for every new project. Deployment Protection Exceptions are also now free on every plan. Exceptions let you make specific preview domains public while the project’s other deployments remain protected. Learn more in the Deployment Protection documentation. Read more

Deployment step now 10% faster

September 8, 2026

The deployment step is now about 10% faster, saving one second on average. Large applications can save up to 12 seconds. Previously, Vercel uploaded a separate routing metadata file for each function path. That metadata is now combined into a single manifest and uploaded once. This improvement applies automatically to all builds. No changes are required. Learn more about builds in the documentation. Read more

GPT Image 2.5 Flare and Sunburst now available on AI Gateway

September 8, 2026

GPT Image 2.5 Flare and GPT Image 2.5 Sunburst from OpenAI are now available on AI Gateway as openai/gpt-image-2.5-flare and openai/gpt-image-2.5-sunburst. Both models generate and edit images with more natural lighting and textures, follow complex visual instructions, and handle detailed layouts and transparent backgrounds. They accept reference images for targeted edits while preserving surrounding elements. Flare is optimized for faster generation and iteration. Sunburst takes longer to generate and offers tighter control for detailed creative work. Use openai/gpt-image-2.5-flare as the model name for faster generation: Change the model name to openai/gpt-image-2.5-sunburst when the result needs more precise control over composition and detail. For targeted editing, pass an image alongside the instruction. This example uses Sunburst: Try Flare or Sunburst in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with

Vercel Sandbox routing is now 18x faster globally

September 8, 2026

Requests to Vercel Sandbox public domains are now routed 18x faster. Domains created with the sandbox.domain() SDK call are now resolved from the nearest regional replica, instead of a single centralized store. Incoming requests reach the process running in the sandbox with less latency. Median domain lookup latency dropped from 62ms to 3.4ms (18x faster). The improvement is largest in regions farthest from the previous store: lookups are now up to 112x faster in Sydney (syd1) and up to 146x faster in Cape Town (cpt1), at p99. This applies automatically to every request to a sandbox domain. There are no pricing changes. Learn more about Vercel Sandbox in our documentation. Read more

Introducing Flat Rate CDN

September 8, 2026

Over the past year, we have improved Vercel’s Pro pricing to better reflect how teams build, with expanded user roles, flexible credits, access to enterprise features, and more. But one area our customers continued to give us feedback on was CDN pricing. CDN costs should be predictable, but viral launches, traffic bursts, or misconfigured routes sometimes turned an otherwise normal month into an unexpected bill. For teams building on Vercel, that uncertainty made growth feel like a problem at exactly the moment it should feel exciting. Today, we are introducing Flat Rate CDN for Pro teams: a new way to pay for CDN usage with a predictable monthly bill, built-in spike protection, and service tiers designed for different traffic profiles. What is Flat Rate CDN? Flat Rate CDN is an alternative to usage-based CDN billing that provides you with a predictable monthly bill. The model is designed around a simple idea: teams should be able to choose the CDN capacity that corresponds to their typical usage profile, pay a known amount monthly, and avoid surprise overages from temporary spikes. Flat Rate CDN has several advantages over usage-based billing: Fixed monthly price: Pay a fixed amou

Flat Rate CDN is now GA for Pro teams

September 8, 2026

Flat Rate CDN is now generally available for Pro teams. Flat Rate CDN is an alternative to usage-based CDN billing that provides you with a fixed monthly bill. Monthly Flat Rate CDN capacity includes: Fast Data Transfer Blob Data Transfer CDN Requests Observability events generated by CDN requests Capacity is set at the team level, not the project level, so a spike on a single project draws on the team-level allowance. Spike protection is on by default. With spike protection, traffic above your tier's capacity is served normally, without degraded performance, and isn't billed (subject to the fair use guidelines). 1M requests and 1 TB of data transfer are included with Pro. Additional monthly capacity can be purchased by upgrading to a higher Flat Rate CDN tier. Get started New Pro teams have Flat Rate CDN on by default, and existing teams can switch over in Billing settings. Read the Flat Rate CDN announcement and learn more about pricing in the documentation. Read more

Ling 3.0 Flash Sante is now available on AI Gateway for free

September 4, 2026

Ling 3.0 Flash Sante from inclusionAI is now available on AI Gateway, free to use through October 4. Ling 3.0 Flash Sante is a health and medicine-focused version of Ling 3.0 Flash. It is a Mixture-of-Experts model with 124B total parameters and about 5.1B active per token, a 256K token context window, and function calling. The model is built for medical reasoning, professional healthcare tasks, deep research, evidence-based retrieval, and multi-step medical workflows. It retains the base model's general reasoning, coding, and agentic capabilities. How to use the model during the free period The standard model ID, inclusionai/ling-3.0-flash-sante, is free through October 4 and begins billing when the offer ends. The free model ID, inclusionai/ling-3.0-flash-sante-free, stops serving when the offer ends instead of billing. Free requests still appear in your spend dashboard and carry a trace, they just cost nothing. To use Ling 3.0 Flash Sante, set model in the AI SDK: Use inclusionai/ling-3.0-flash-sante-free if you want the model to stop serving when the offer ends rather than start billing. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding

GPT 6 Astra now available on Vercel AI Gateway

September 4, 2026

GPT 6 Astra from OpenAI is now available on AI Gateway. It is designed for long-running agentic work across software engineering, computer and browser use, scientific research, and professional workflows. Astra can navigate software, complete forms, organize records, analyze data, run simulations, and build and test websites. It also improves instruction following during evolving tasks. It can incorporate new requirements without losing earlier constraints, continue independent work while waiting for clarification, and use reasonable assumptions. To call GPT 6 Astra in code or in coding agent integrations, use openai/gpt-6-astra: To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect Codex, Cursor, and more, and select openai/gpt-6-astra in the agent. Try GPT 6 Astra in the model playground. Read more

Cursor Cloud Agents can now run in Vercel Sandbox

September 3, 2026

Cursor Cloud Agents can now run in Vercel Sandbox instead of Cursor's hosted machines. Cursor manages the agent harness and inference loop. Its Self-Hosted Machines APIs let you supply the execution environment where agents clone repositories, edit files, and run commands and tests. Self-Hosted Machines requires a Cursor Enterprise plan. Vercel Sandbox provides that execution environment as an isolated Firecracker microVM for each agent request. Vercel Functions and Vercel Workflow form a durable control plane that claims queued agent requests, provisions workers, monitors sessions, and cleans up automatically. With this architecture, you get: A scale-to-zero worker pool without long-lived virtual machines A dedicated, isolated Sandbox for every agent request Durable retries when a worker or session fails Short-lived, user-scoped credentials inside each Sandbox Follow the step-by-step guide to deploy the reference implementation to your own Vercel account, or learn more in the Vercel Sandbox documentation. Read more

Basic build machines are now available on Pro and Enterprise

September 3, 2026

Pro and Enterprise teams can now select Basic build machines. Basic build machines have 2 vCPUs and 8 GB of memory, offering a more cost-efficient option for smaller apps or agents that build with fewer resources. New Pro and Enterprise projects still default to Elastic build machines, which scale automatically and are recommended for most use cases. To use Basic build machines, update your project configuration in team settings, project settings, or with Vercel CLI 59.6.0 or later: Hobby accounts continue to build on the same 2 vCPU machines, now called Basic build machines. Basic build machines are charged at the same $0.0035 per vCPU per minute, which works out to $0.007 per build minute. See the pricing page or read the build machines documentation. Read more

Free domain with Pro offer now includes .app and .dev

September 2, 2026

We recently announced that all Pro teams can claim a free domain for one year. That offer now includes .app and .dev, bringing the eligible TLDs to eight: .app .dev .online .site .space .store .tech .website The offer still covers one domain per team, and premium domains are not eligible. After the first year, your domain will renew automatically at the regular price shown at checkout when you claim it. Claim yours at vercel.com/domains, and see the free domain FAQ for full terms. Read more

Muse Spark 1.3 now available on AI Gateway

September 2, 2026

Muse Spark 1.3 from Meta is now available on AI Gateway, in both the standard and contributor pricing tiers. This model improves on prior Muse Spark models at agent work and coding, with a 1M token context window and text, image, and PDF input. On coding it takes fewer turns and writes less filler than the previous release. To use Muse Spark 1.3, set model to meta/muse-spark-1.3: Contributor tier Muse Spark 1.3 Contributor is a pricing tier on the same model rather than a separate one, with the same weights, capabilities, and context window. The difference is that Meta uses the inputs and outputs sent to this tier to train and improve its models, and pricing is lower in exchange. Model Input Output Cached input meta/muse-spark-1.3 $1.25 $4.25 $0.15 meta/muse-spark-1.3-contributor $0.10 $0.20 $0.002 Rates are per million tokens and unchanged from Muse Spark 1.2 on both tiers. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect Claude Code, Codex, Cursor, and more, then select meta/muse-spark-1.3 in the agent. Try Muse Spark 1.3 in the model playground. You can view all language models available on AI Gateway. Read more

GLM-5.3 is 50% off through DigitalOcean on AI Gateway

September 2, 2026

GLM-5.3 is 50% off on AI Gateway through Tuesday, September 8, in partnership with DigitalOcean. How to use the model during the offer period Using the promo name (zai/glm-5.3-promo-50) gets the discounted rate. It routes only to DigitalOcean, with no fallback to another provider, and it stops serving when the offer ends. Using the standard name (i.e., zai/glm-5.3) with provider options to sort DigitalOcean as the preferred provider keeps working after September 8 and routes across every provider that serves the model, at their usual rates. Because the promo name goes away when the offer ends, treat it as something you switch on for the window rather than hardcode. To keep the standard name in your code instead, pin the provider with order: ['digitalocean'] under providerOptions.gateway, which prefers DigitalOcean and falls back to the others if it cannot serve the request. GLM-5.3 takes text input, with a 1M token context window and a maximum output of 128K tokens. Discounted requests appear in your spend dashboard and carry a trace like any other request. Try GLM-5.3 in the model playground. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway codi

Gemini 3.8 Flash now available on AI Gateway

September 2, 2026

Gemini 3.8 Flash from Google is now available on AI Gateway. The model is 50% off through December 31st. It has a 1M token context window, accepts text, image, PDF, and video input, returns text, and supports tool calling and web search. Maximum output is 65,536 tokens. Gemini 3.8 Flash improves on prior Flash models at software engineering, agent work, and multi-step reasoning, at the same speed and cost as the previous release. Thinking is on by default. To use Gemini 3.8 Flash, set model to google/gemini-3.8-flash: To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect agents like Claude Code, OpenCode, Cursor, Pi, and more and select google/gemini-3.8-flash inside the agent. Try Gemini 3.8 Flash in the model playground. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. You can view all language models available on AI Gateway. Read more

Qwen 3.8 Max 0902 now available on AI Gateway

September 1, 2026

Qwen 3.8 Max 0902 now available on AI Gateway Qwen 3.8 Max 0902 from Alibaba is now available on AI Gateway. This is a new snapshot of Qwen 3.8 Max, with the gains concentrated in coding on larger projects, long-horizon work that runs without supervision, and agent runs. Vision handling is more accurate on charts and dense documents. To use Qwen 3.8 Max 0902, set model to alibaba/qwen3.8-max-0902: The dated ID pins this snapshot, so a later release will not change what your requests run against. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect agents like Claude Code, Codex, OpenCode, Cursor, Pi, and more and select alibaba/qwen3.8-max-0902. Try Qwen3.8-Max-0902 in the model playground. You can view all language models available on AI Gateway. Read more

Claude Fable 5.1 now available on AI Gateway

September 1, 2026

Claude Fable 5.1 from Anthropic is now available on AI Gateway. Fable 5.1 improvements compared to previous Claude models are concentrated in long, multi-stage work like agentic coding, knowledge work, and research that takes several rounds of searching and following up. Anthropic ships Fable 5.1 with cybersecurity and biology safety classifiers enabled. Finding vulnerabilities in source code is allowed, but some routine coding and debugging may still be refused. To ensure requests are still serviced when the safety classifiers are triggered, use model fallbacks. Add a models array to providerOptions.gateway listing the models to try. AI Gateway sends the request to Fable 5.1 first, and if Anthropic refuses it, works down the array in order and returns the response from the first model that succeeds: This request falls back to Opus 5, then Sonnet 5, if a safety classifier is triggered. The same models option works on every AI Gateway API format, including Chat Completions, Messages, and OpenAI Responses. Try Fable 5.1 in the model playground. To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect agents like Claude Code,

Compute that takes any shape

September 1, 2026

Workloads differ in how long they run, how much memory they take, and how much of the environment they control. That previously meant a different compute primitive for each job, so product iterations required provisioning infrastructure, not just working on a feature. We built Vercel so that none of this has to be your problem. The compute layer is a single system with a simple job. Take any workload, assemble the machine it needs, swap configuration on the fly, and absorb burst capacity in real time. We call this Fluid. Builds ran on it first, then sandboxes, and now functions. If you've shipped on Vercel, you've been on Fluid without knowing it. Fluid compute now runs over 15 million builds a day, 25 million sandboxes a week, and a trillion requests a month. A history of compute Changing a machine used to mean changing it by hand. You went down to the computer store, bought a hard drive or a stick of RAM, and swapped it in yourself. Then you could rent bare metal when you needed it, and use it to serve a website, host a database, or send email. Then the cloud let you request a machine in any configuration you wanted, choose its operating system, use it, and throw it away when you

AWS PrivateLink is now available on Pro and Enterprise

September 1, 2026

AWS PrivateLink is now available for Pro and Enterprise teams as part of Advanced Networking. It allows Vercel Functions and builds to connect to AWS-hosted services without sending traffic over the public internet. Use it to reach AWS-hosted databases such as RDS, Aurora, and Neon; SaaS services such as Snowflake and MongoDB Atlas; internal services behind an AWS Network Load Balancer; and S3 or DynamoDB through gateway endpoints. To configure PrivateLink, open your project's Networking settings and enable Advanced Networking. Select New Connection, then enter the service name and region. The service must accept connections from all AWS principals or allowlist the IAM role Vercel provides for your team. Vercel creates the connection and provides a stable hostname. Use this hostname from your Functions and builds to reach the service privately. Pricing: The first AWS PrivateLink connection is included with Advanced Networking. Each additional connection costs $30 per month. PrivateLink data transfer costs $0.04 per GB. Learn more in the AWS PrivateLink documentation. Read more

fx is now available in the AI SDK harness layer

August 31, 2026

The AI SDK harness layer now supports fx, Vercel's lightweight, open-source coding agent. The harness layer provides one API for running coding agents in your application, so you can add fx without building a separate integration. Configure HarnessAgent with the official @ai-sdk/harness-fx adapter: The @ai-sdk/harness-fx adapter connects to fx over the Agent Client Protocol (ACP) using @ai-sdk/harness-acp. fx joins Claude Code, Cline, Codex, Cursor, Deep Agents, Grok Build, OpenCode, and Pi in the list of supported harnesses. Read the fx harness documentation to get started. Read more

Set per-user budgets on AI Gateway

August 31, 2026

You can now use AI Gateway budgets to set a dollar spending limit for each user on your team. The limit covers spend from every API key attributed to the user, along with their app tokens. After it's reached, AI Gateway rejects new requests until the budget resets or is increased. This is useful for controlling spend from coding agents and other workloads that run without supervision, and prevents one user from consuming all of the team's shared budget. Set user budgets Open the Users view on the Budgets page to set two kinds of user budgets: Default budget: If set, applies separately to every person without a custom budget, including users added later. Each user gets their own allowance rather than sharing one limit across the team. Custom budget: Applies to one person and overrides the default. Removing it returns that user to the default budget. Both default and custom budgets reset monthly by default, but you can change them to reset daily, weekly, or not at all. Budgets also support usage alerting via email to notify when you are at 50%, 75%, and/or 100% of your allocation. User budgets don't replace API key, project, or team budgets. A request must remain within every budget

How our agents build on-brand pages with design.md

August 31, 2026

Across Vercel, we use coding agents to design and build pages that have to look and feel like Vercel. The typography, color, and composition all need to carry the same judgment that we put into the pages we already ship ourselves. We recently wrote about product-design, our skill that teaches agents how we design when they work in our codebases. The skill lives in each repository alongside the code that it governs, explaining how agents can find and understand our design system as well as product guidelines for whatever they are building. This works great when agents are working in our codebases, where everything the skill needs is right there. But what about reports, proposals, and the one-off pages that still have to look like Vercel but get made in tools that can't read any of those files? For us, the answer was design.md, one public file any agent can load. How we approached building design.md What made product-design work well was that the design system and product guidelines were sitting right there in the repository for agents to read. We needed a way for agents and tools outside that environment to reach the same knowledge, so that the pages coming out the other end would s

Vercel Sandbox now calculates snapshot storage costs daily

August 31, 2026

Vercel now calculates charges for Sandbox snapshot storage from each day’s average usage and adds them to your monthly invoice. Previously, billing used one average across the entire monthly billing period. You can track these daily charges on the team’s Usage page and see how changes in storage affect your cost. The rate remains $0.08 per GB-month; only how usage is calculated has changed. The update takes effect automatically at the start of the next billing period for Pro and Enterprise teams. No action is required. Learn more about Sandbox snapshot storage in the documentation. Read more

MiniMax H3 and H3 Max are 50% off on AI Gateway

August 30, 2026

MiniMax H3 and H3 Max are 50% off on AI Gateway from August 30 through September 13, in partnership with MiniMax. The discount covers requests billed through AI Gateway, at every duration and in every aspect ratio the model supports. H3 generates 2K video from a text prompt, a starting image, a pair of first and last frames, or reference images, video, and audio. H3 Max trades resolution for speed: it renders faster at 480p and 768p, and it takes a text prompt or a starting image. The model IDs (minimax/minimax-h3 and minimax/minimax-h3-max) are unchanged, so requests you already send pick up the discounted rate with no code change: Renders take minutes, so poll runs the generation as a background job and makes short status requests until it lands, rather than holding one long request open. See asynchronous generation for the webhook and start-and-status routes. Get started Create an API key in the AI Gateway section of your dashboard, or generate a clip in the browser first from the model playground. Current rates for every model are on the pricing page. You can view all video models available on AI Gateway. Read more

Hy4 Preview now available on AI Gateway

August 28, 2026

Hy4 Preview from Tencent is now available on AI Gateway. Hy4 Preview is an open-source Mixture-of-Experts model with 770B total parameters and 49B active per token, aimed at long-horizon coding, document analysis, game development, and scientific reasoning. It serves a context window of 1M tokens. To use Hy4 Preview, set model to tencent/hy4-preview in the AI SDK: To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect agents like Claude Code, Codex, OpenCode, Cursor, Pi, and more and select tencent/hy4-preview inside the agent. Try Hy4 Preview in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. You can view all language models available on AI Gateway. Read more

Build and deploy eve agents from the Vercel dashboard

August 28, 2026

You can now create eve agents directly from the Vercel dashboard, giving you a deployed agent that you can chat with right away. The builder scaffolds the agent, creates a private Git repository with its files, and deploys it as a new Vercel project on your team. In a few clicks, you get a working agent, backed by a repo with code you can use to customize it. From the builder you can: Write the instructions that define the agent's identity Pick any model available through AI Gateway Add a Next.js web chat or Slack channel where people can talk to the agent Give the agent tools and data from other services through built-in connections like Linear and Notion, or a custom MCP server Select Add New → Agent in your dashboard to create your first agent. Read more

Vercel CLI expands commands for DNS, domains, and projects

August 28, 2026

The Vercel CLI now provides dedicated commands for managing DNS records, domains, and projects. This brings more of the functionality available through the Vercel dashboard and API to the terminal, where it can be used interactively, in scripts, or by agents. Inspect and update DNS records Retrieve the complete configuration for a DNS record or update it in place using its record ID. vercel dns update supports changes to the record name, type, value, TTL, MX priority, and comment. SRV records can also be updated using their priority, weight, port, and target fields. Renew domains bought on Vercel Renew a domain on demand or control whether it renews automatically. The renewal command displays the current renewal price and asks for confirmation before completing the purchase. Control project state and observability Pause or resume a project and configure its observability features directly from the CLI. Web Analytics and Speed Insights could already be enabled from the CLI; they can now be disabled as well. Add and remove project members Add a project member with a specific role or remove an existing member. All of the new commands support structured JSON output for scripts and agen

Run Claude Managed Agents with Chat SDK

August 28, 2026

You can now run Claude Managed Agents with Chat SDK. Claude Managed Agents handles the agent loop server-side, including the model, tools, session state, and sandboxed web research. That means you can ship a Slack research bot built on Claude Managed Agents and Chat SDK, with one persistent session per thread and streamed briefs with sources. What you get Token-by-token streaming: Replies render as the model writes them, over a single streamed response. Live activity feed: Tool calls and model requests are available during the turn, so you can surface a trace in the chat. No database to run: The Managed Agents session stores the conversation, so the sidebar, transcript, and replay read from it, no server-side state of your own. Portability by design: Swapping a few lines in the handler moves your agent to Teams, Google Chat, Discord, WhatsApp, and 30+ other platforms. Get started To try Claude Managed Agents with Chat SDK: Deploy a Slack research agent on Vercel in one click Follow the step-by-step guide or view the template source Read more

Cursor is now available in the AI SDK harness layer

August 27, 2026

The AI SDK harness layer now supports Cursor through the official @ai-sdk/harness-cursor adapter. The harness layer lets your application run different coding agents through the same HarnessAgent interface, so you can switch agents without changing your application code. Pass cursor to HarnessAgent: Under the hood, the adapter uses @ai-sdk/harness-acp to connect Cursor to HarnessAgent through the Agent Client Protocol (ACP). Supported harnesses now include, in addition to Cursor, Claude Code, Cline, Codex, Deep Agents, Grok Build, OpenCode, and Pi, with more coming soon. Read the Cursor harness documentation to get started. Read more

Find deployments faster with redesigned filters

August 27, 2026

The Deployments page now has redesigned filters that make it faster to find a specific deployment. You can: Apply suggested filters for common searches. Type to find an option instead of scrolling through the list. Describe what you’re looking for in natural language, and Vercel automatically applies the matching filters. Open any project's Deployments page to try the new filters. Read more

The best workflow engine is a programming language

August 27, 2026

The idea of orchestrating long-running, stateful logic on top of unreliable, stateless infrastructure isn't new. We've had message queues, job runners, microservice choreographies, and full-blown workflow engines for a long time. What we never had was a version of it that felt good to write. I'd spent about six months working on a fork of Temporal, mostly on weekends, trying to turn it into a serverless answer with DX that felt more Vercel-native. Eventually it dawned on me that to ship that experience, I'd need to own the execution environment too. So I dropped the fork, joined Vercel, and started hacking on a new framework from scratch alongside Nathan Rajlich. Code is already a DAG A workflow is a DAG, a directed acyclic graph. Before Temporal (and Cadence before it), nearly every workflow framework made you draw that DAG by hand. Apache Airflow is the canonical example. You describe your pipeline as an explicit graph of tasks and dependencies, and your actual logic gets buried inside the nodes." That always felt backwards to me. We already have a tool for expressing "do this, then that, and these in parallel, and branch here." It's called a programming language. An abstract syn

Ling 3.0 Flash Fin now available on AI Gateway for free

August 27, 2026

Ling 3.0 Flash Fin from Inclusion AI is now available on AI Gateway and free to use through September 25. Ling 3.0 Flash Fin is a finance-focused version of Ling 3.0 Flash. It has a 256K-token context window, produces up to 32K output tokens, and supports reasoning and function calling. The model is designed for financial research and analysis, including multi-step workflows that user multiple tool calls before producing an answer. Choose a model ID based on what should happen after the free period: Continue after September 25: Use inclusionai/ling-3.0-flash-fin. Requests are free during the offer and begin billing at the standard rate when it ends. Stop after September 25: Use inclusionai/ling-3.0-flash-fin-free. This model ID returns an error after the offer ends, preventing future charges. To use Ling 3.0 Flash Fin with a coding agent, run vercel ai-gateway coding-agents setup, then select either model ID as your agent’s model. See the coding agents guide for more. Try Ling 3.0 Flash Fin in the model playground, or browse all language models. Read more

Muse Image now available on AI Gateway

August 26, 2026

Muse Image from Meta Superintelligence Labs is now available on AI Gateway. It is their first image model and a separate family from Muse Spark, returning images rather than text. Send a prompt and get an image back, or send an image with an instruction and get it changed. One model does both, so you don't switch models to move from generating to editing. To use Muse Image, set model to meta/muse-image-1.0 and call generateImage from the AI SDK: To steer the result toward art you already have, pass reference images in prompt.images alongside the text, and the model blends them into what it draws. Editing Pass the image you want changed in prompt.images with an instruction, and the model changes what you asked for and leaves the rest: Try Muse Image in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. You can view all image model

Gemini 3.5 Transcribe now available on AI Gateway

August 26, 2026

Gemini 3.5 Transcribe from Google is now available on AI Gateway for recorded and live audio: google/gemini-3.5-transcribe transcribes a complete audio file in one request. google/gemini-3.5-transcribe-live transcribes audio over a WebSocket and returns text as the audio arrives. Both models automatically detect more than 85 languages, including when a speaker switches languages. You can also provide custom vocabulary to improve the transcription of names, technical terms, and uncommon spellings. The model for complete recordings can also identify speakers and return word-level timestamps. Streaming transcription is available in AI SDK 7. Install the latest AI SDK and AI Gateway provider: Transcribe live audio Use streamTranscribe with a ReadableStream of raw audio chunks. Set inputAudioFormat to match the audio being sent: Transcribe a complete recording Use transcribe to send a complete audio file and receive the finished transcript: Try Gemini 3.5 Transcribe Live in the model playground, browse all transcription models, or read the speech quickstart. Read more

Qwen 3.8 Flash now available on AI Gateway

August 26, 2026

Qwen 3.8 Flash from Alibaba is now available on AI Gateway. It takes text and images as input, serves a context window of 1 million tokens, and can return up to 65k tokens in a response. Alibaba recommends it for coding, tool use, and multi-step agent workflows. To use Qwen3.8-Flash, set model to alibaba/qwen3.8-flash in the AI SDK: To use it in a coding agent, see the coding agents guide, then run vercel ai-gateway coding-agents setup to connect agents like Claude Code, Codex, OpenCode, Cursor, Pi, and more and select alibaba/qwen3.8-flash inside the agent. Try Qwen3.8-Flash in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Read more

GLM 5.3 Flash now available on AI Gateway

August 26, 2026

GLM 5.3 Flash from Z.ai is now available on AI Gateway as zai/glm-5.3-flash. GLM 5.3 Flash is a multimodal coding model with a 1M-token context window. It accepts both text and image inputs and supports function calling, structured output, and streaming. To include images, pass them with text in a message. A request can include multiple images using URLs, Base64 data URLs, or binary data: To use GLM 5.3 Flash with a coding agent, run vercel ai-gateway coding-agents setup, then select zai/glm-5.3-flash as your agent's model. Try GLM 5.3 Flash in the model playground, or browse all language models. Read more

Vercel Security Dashboard is now generally available

August 26, 2026

The Vercel Security Dashboard is now generally available on all plans, giving you one place to see your security posture across every account and project. You can access the Security Dashboard in the UI or run vercel security check in the Vercel CLI. As teams grow and coding agents make it faster to spin up projects, small misconfigurations add up quietly. The Security Dashboard automatically flags issues like: Team members without 2FA Long-lived credentials that can be replaced with OIDC Public preview deployments Non-sensitive and stale environment variables The Security Dashboard UI Misconfigurations are ordered by risk with the most severe first, and each finding links to the settings that fix them. Findings that are just noise for your team can also be muted, and all results can be exported to a CSV file for triage or reporting. The Security Dashboard CLI You can also run the same checks in your terminal through the new Vercel CLI vercel security check command, allowing your agents to work through the findings for you. An agent using vercel security check --findings can read what failed, apply the fix and re-check to confirm. Fixes an agent can apply include: Turning on Git fo

Python projects now support routing rules

August 26, 2026

Python projects can now use routing rules to set response headers or rewrite requests to internal paths, including apps built with FastAPI, Django, and Flask. The Vercel CDN evaluates rules before requests reach your application, so changes apply without a new deployment. For example, this FastAPI app serves a /new route: To send requests for /old to that route, create a rewrite from the CDN tab in your project dashboard or with the Vercel CLI: Published rules take effect immediately across all regions, and you can roll back to a previous version from the History tab. You can also manage rules from Python with the vercel SDK, which wraps the Vercel REST API. add_route stages the rule, and update_route_version publishes it: Rewrites can also be defined in vercel.json: Project-level rules live outside your deployment and are evaluated before its routes, so a published rule takes precedence over a vercel.json rewrite for the same path. Each surface fits a different workflow: How you define it Takes effect Best for CDN tab in the dashboard On publish Editing and testing rules visually Vercel CLI On publish One-off changes from the terminal vercel Python SDK or REST API On publish Autom

AI Gateway now supports asynchronous video generation

August 25, 2026

Video generation on AI Gateway can now run asynchronously. By default, generateVideo keeps one HTTP request to AI Gateway open until the result is ready. Because video generation can take seconds or minutes, that request can exceed request timeouts. With asynchronous generation, your application can receive a webhook, poll for completion, or start a generation and retrieve the result in a later request. Choose an option based on whether your process can keep running and whether your application can receive webhooks: Existing generateVideo calls continue to work as before. All four options support text-to-video, image-to-video, reference-to-video, and other video inputs. Upgrade the SDK Install the latest versions of the AI SDK and AI Gateway provider: Use asynchronous video generation Wait for completion in a Workflow An easy way to consume the completion webhook is a Workflow SDK. The workflow creates its own webhook URL, passes it to startVideo, and suspends until AI Gateway delivers the completion event. Install the Workflow SDK alongside the AI SDK: While the video renders, the workflow run is suspended and resumes when AI Gateway delivers the terminal event. Use a webhook with

Vercel Connect is now generally available

August 25, 2026

Vercel Connect is now generally available on all plans and in v0. Instead of storing long-lived provider secrets, your code requests short-lived, scoped tokens at runtime. Deployments authenticate with their existing Vercel OIDC identity. Each token is scoped to the task, refreshed automatically, and expires on its own. Any service with one command Register a connector once from the CLI. Pass the service name and the CLI pre-populates the brand name, icon, auth type, and MCP or discovery URL, then prompts for any credentials the service requires: Connect ships with 100+ preset connectors for tools like Notion and Workday, managed connectors for Slack, GitHub, Linear, Salesforce, Snowflake, and Microsoft, plus generic OAuth, API key authentication, and MCP servers. Tokens at runtime, not secrets at rest Request a token only when your code needs one, with getToken: Switch the subject from the app to a named user and the token acts on that user's behalf, triggering the authorization flow when consent is needed. Access your team can inspect and prove New at GA: Fine-grained RBAC controls who can create and manage connectors Audit logs record authorization and connector activity Token a

Chat SDK now supports Slack Enterprise Grid

August 25, 2026

Chat SDK's Slack adapter now supports Slack Enterprise Grid. Bots installed org-wide work across every workspace, with correct token resolution, tenant-scoped caches, and event retry deduplication. The adapter now stores org-wide installations by enterprise ID. This matches how tokens are resolved for incoming events, slash commands, and interactive payloads. SlackInstallation records the new identity fields: Token resolution behaves the same over HTTP webhooks and Socket Mode. Events route by the installation identity in the envelope's authorizations field, which keeps routing correct for Slack Connect shared channels. The adapter is also hardened for multi-workspace deployments: User profile and mention caches are scoped per installation, so one tenant's data never resolves for another. API calls made with an org-wide token pass the event's team_id automatically, which Slack requires for workspace-scoped methods. Retried event deliveries are deduplicated for 24 hours, covering Slack's Delayed Events redeliveries. Outgoing mentions accept W-prefixed Grid user IDs alongside U-prefixed IDs. Single-workspace installations are unaffected, and getInstallation and deleteInstallation wor

Vercel Connect now supports Linq

August 25, 2026

Vercel Connect now includes a managed connector for Linq, so your apps and agents can send and receive messages over iMessage, RCS, and SMS. As a Vercel Managed Connector, Vercel can create a Linq account and phone number for you, or link an existing account. You never manage credentials yourself. Create a connector from the dashboard or Vercel CLI: Give your eve agent a phone number The connector powers the new Linq channel in eve. Run eve add channel/linq, choose Vercel Connect, and eve wires up the connector, phone numbers, and webhook for you: The channel marks accepted messages as read and continues the same eve session for every message in a Linq conversation. Out of the box, the channel supports read receipts, typing indicators, and mid-conversation steering for a more natural messaging experience. Use the connector in any project Call getToken with your connector, and the @vercel/connect SDK exchanges your Vercel deployment's OIDC identity for a short-lived Linq token. Pass it as the bearer token on any Linq Partner API request: Linq delivers each message over the best protocol the recipient supports, starting with iMessage, then RCS, then SMS. The same token works across t

Bring your agent to Notion with Chat SDK

August 25, 2026

Your team already works in Notion. Now your agent can too. With the new Notion adapter for Chat SDK, the same agent you run on Slack, Discord, GitHub, Teams, or WhatsApp can join comment discussions on your Notion pages, no separate codebase required. Each Notion page maps to a channel and each comment thread to a thread, so replies stay threaded automatically. The adapter supports mentions, message editing, conversation history, and up to three file attachments. By default, your bot replies when @-mentioned and where mentions aren't available in a workspace, it can trigger on a keyword or on all comments instead. Notion doesn't support buttons, modals, or reactions, and cards render as markdown. Discussions can start at the page or block level, but not on inline text ranges. History fetches return only open comments. Read the documentation to get started, browse the directory, or build your own. Read more

Chat SDK now supports XChat

August 25, 2026

You can now build bots that hold end-to-end encrypted 1:1 and group conversations on XChat with the new XChat adapter for Chat SDK. The adapter handles all encryption, key management, and signature verification automatically. Bots can also message users first, as long as the user has encrypted chat set up and follows the bot. XChat has no markdown rendering, so the adapter falls back automatically: URLs and @mentions render as tappable links, tables as ASCII code blocks, and cards as text with a link preview. Streaming works through message edits. Read the XChat adapter documentation to get started. Read more

Vercel applications are protected from Next.js August 2026 security vulnerabilities

August 25, 2026

Summary Two vulnerabilities affecting Next.js were disclosed in the August 2026 Security Release. Next.js applications hosted on Vercel are protected and require no customer action. Next.js August 2026 vulnerabilities Next.js disclosed the following critical vulnerabilities: GHSA-2xp9-vwfh-vxw4 originates in the upstream libheif dependency and can lead to unauthenticated remote code execution when Image Optimization processes a crafted AVIF input. CVE-2026-75604 (GHSA-p293-qw3h-jr36) can lead to unauthenticated remote code execution on Windows-hosted Next.js servers in applications using the Pages Router and App Router without Cache Components. After the AVIF vulnerability was identified, Vercel applied protections to its managed Image Optimization service. Impact on Vercel deployments Applications hosted on Vercel are protected. No upgrades, configuration changes, or redeploys are required. Once the AVIF vulnerability was identified, Vercel disabled AVIF optimization across its managed Image Optimization service. AVIF inputs are served as-is and do not pass through the affected processing path. The second vulnerability only affects servers using a Windows filesystem. Vercel's Next

Speed Insights now has a free tier

August 25, 2026

Speed Insights now has a free tier that gives you a high-level performance overview from your real users. The new free tier: Is available on every plan, for any number of projects Includes 10,000 events per team, every 30 days Install Speed Insights: Previously, free Speed Insights was limited to a single project on Hobby, and upgrading to Pro meant losing access unless you paid for the add-on. Now your free tier carries over, and you only pay if you upgrade. The paid product, now called Speed Insights Plus, includes deeper diagnostics and historical data. If you were already paying for Speed Insights, only the name has changed, and your capabilities and price are the same. Install @vercel/speed-insights in your project to start reporting real-user data, or open the Speed Insights dashboard to view your results. Learn more in the Speed Insights documentation. Speed Insights reports how real visitors experience your pages, using live traffic rather than lab conditions. Speed Insights Plus adds slow-route diagnosis and detailed breakdowns for finding where pages get slow, and historical data to track performance over time. It also provides access to Drains and the Vercel CLI for work

Introducing Run SDK: secure eval for your agents

August 25, 2026

Agents increasingly write TypeScript programs to coordinate tools and process their results. Once those programs touch real applications, some steps require authentication, while others need human approval. Executing that code with eval gives it the same access as the application around it, including its secrets and internal services, and leaves no durable way to pause at those boundaries. Today, we're releasing the Run SDK, a package for executing untrusted JavaScript and TypeScript without giving it direct access to your application or system. Applications expose narrow host functions and can interrupt execution for authentication or human-in-the-loop approval. The program resumes after a decision without repeating completed work. A small interface to the host The Run SDK evaluates JavaScript or type-stripped TypeScript in a fresh QuickJS context inside a worker thread, with no direct route to Node.js or the network. The application exposes selected operations through hostFunctions. These are regular functions that become callable globals inside the sandbox: Here, the generated program knows about store.listOrders(). The database client and its credentials remain in the applicati

The end of credential sprawl for agents

August 25, 2026

Every useful agent reaches beyond your codebase. It posts to Slack, opens pull requests, queries Snowflake, or calls an internal API. That reach is what makes it valuable, and it's also where the risk lives, because for years, granting it meant provisioning a long-lived token and hoping it never leaked. Vercel Connect replaces long-lived tokens with ones your code requests at runtime, scoped to the task and expiring on their own. During the public beta, we've grown the ecosystem past 100 connectors, unified how they work, and added the governance capabilities teams need in production. Today, Vercel Connect is generally available. Vaults don't fix long-lived tokens Managing credentials has become its own workload. Teams write rotation scripts, copy secrets across environments, and share tokens between users. Putting a token in a vault made it harder to steal, but no less dangerous once stolen. It never expires, and no vault limits what a leaked credential can do. Agents compounded the problem by touching more systems with greater autonomy, more often. Yet the tools to contain a secret haven't changed. With Vercel Connect, your app never stores credentials. It requests one: You regis

MiniMax M3 and M2.7 are free on AI Gateway

August 25, 2026

MiniMax M3 and M2.7 are free on AI Gateway via GMI Cloud through Sunday, September 6. Use minimax/minimax-m3-free or minimax/minimax-m2.7-free to route requests to GMI Cloud. These model IDs will return an error after the free period ends. To keep requests working after the free period, use the standard model ID without the -free suffix and place GMI Cloud first in the provider order: AI Gateway tries GMI Cloud first and can fall back to another provider if GMI Cloud can't serve the request. This lets the same code continue working after the free period. Outside the free GMI Cloud route, requests are billed at the serving provider's rate. Both routing options work across all AI Gateway APIs. To use a free MiniMax model with a coding agent, run vercel ai-gateway coding-agents setup, then select one of the -free model IDs in your agent. Try MiniMax M3 and M2.7 in the gateway playground. Read more

Wan 3.0 now available on AI Gateway

August 25, 2026

Wan 3.0 from Alibaba is now available on AI Gateway as alibaba/wan-v3.0-video. Wan 3.0 combines text-to-video, image-to-video, first- and last-frame conditioning, and reference-based generation in one model. References can include images, video, and audio. It generates clips up to 30 seconds at 30 fps in 480p, 720p, or 1080p, with synchronized audio. Previously, Wan 2.7 required separate -t2v and -r2v model IDs and was limited to 15-second clips at 24 fps. Generate a video Wan 3.0 supports asynchronous generation, so no HTTP request needs to remain open for the entire render. Pass a webhook to receive an event when the generation finishes: Learn about asynchronous generation options in the docs, including how to verify webhook deliveries. Add references Pass image, video, or audio references through inputReferences, including the source and media type for each one. Images accept hosted URLs or base64. Video and audio references require hosted URLs. First- and last-frame conditioning accepts one image for each frame and can't be combined with other references. Try Wan 3.0 in the model playground, or browse all video models. Read more

Vercel Sandbox is now globally available

August 24, 2026

Vercel Sandbox now runs globally, starting with four regions: iad1 (Washington, D.C.), sfo1 (San Francisco), cle1 (Cleveland), and cdg1 (Paris). iad1 remains the default. Support for all Vercel regions is coming soon. Choose a region close to the databases, object storage, and other services your sandboxes access to reduce latency. Region selection is available on all plans. Pro and Enterprise teams can also configure failover regions. If the primary region is unavailable, new sandboxes start in the closest configured failover region. Failover only applies to regions you configure; without it, a sandbox starts in its selected region, or the project default. Update the SDK or CLI you use to the latest version: Configure the default region and failover regions from Settings > Sandboxes in your project, or with the Vercel CLI: Those defaults apply to every sandbox. To override them for a single sandbox, pass region and optionally failoverRegions at create time with the SDK: Or with the CLI: Learn more in the Sandbox regions documentation. Read more

Environment variables now use Config and Secret types

August 24, 2026

When you add or edit an environment variable in Vercel, you now choose Config or Secret instead of using the Sensitive toggle. Existing variables marked Sensitive are automatically treated as Secrets and continue to work without migration. Config: The value remains readable after saving for members with access. Use Config for non-sensitive values you may need to inspect later, such as variables with a public framework prefix. Secret: The value remains available to your deployments and can be replaced, but members cannot view or retrieve it after saving. Use Secret for passwords, API keys, and tokens. You can select an environment or Preview branch for each value. The environment variable list in the dashboard shows each variable’s type and where it applies. Team policy changes The Enforce Sensitive Environment Variables team policy is deprecated with this update. When enabled, it required every environment variable created by a team member to be Sensitive, including non-sensitive configuration. With Config and Secret types, members can choose the appropriate type for each variable. A new Separate Production Secret Values policy is available in your Security settings. When enabled,

Bun runtime now supports large functions and extended max duration

August 24, 2026

The Bun runtime on Vercel Functions now supports larger package sizes up to 5GB uncompressed and extended max duration for up to 30 minutes, two betas that previously ran on Node.js and Python only. Large functions raise the standard 250MB package size limit to 5GB, and extended max duration raises the generally available 800-second ceiling to 1800 seconds for Pro and Enterprise teams. Both features require Fluid compute to be enabled. New projects are enrolled in the large functions beta automatically, while existing projects created before July 2026 opt in by adding VERCEL_SUPPORT_LARGE_FUNCTIONS=1 as an environment variable in your project settings, then redeploying. Durations above 800 seconds are set per function in vercel.json rather than as a project-level default: The server.ts entry point serves the long-running route with Bun: During the beta, these features are unavailable on projects using Secure Compute or Static IPs. Read the large functions and extended max duration documentation. Read more

Connect v0 apps to Slack, Google, and 100+ other services

August 24, 2026

Apps and agents built in v0 can now securely connect to more than 100 services, including Slack, Google, Notion, GitHub, and Salesforce, through Vercel Connect. Ask v0 to connect your app to a service. It opens the connector setup and guides you through it. You set up each connector once for your team, then reuse it across apps. For managed connectors like Slack and GitHub, Vercel handles app registration, so you don’t need to create or configure an app with the provider. For other services, you can provide your own credentials. In both cases, Vercel Connect issues short-lived tokens at runtime, so your application doesn’t need to store or rotate long-lived secrets. Using v0 and Vercel Connect, Danny from KERNEL built a voice-controlled browser agent that generates Playwright code to operate a real browser. We built a personal dashboard that brings Slack messages, Gmail, and Linear issues into one view and summarizes what needs attention. Read the documentation or try it yourself in v0: Read more

Elastic build machines now use Turborepo cache hits to prevent downgrades

August 24, 2026

Elastic build machines now consider Turborepo cache hits when deciding whether to use a smaller build machine. A warm-cache build no longer triggers a downgrade. A warm-cache build can use less CPU and memory than the same build with a cold cache. Downgrading based on that lower usage could leave a later cold-cache build without enough resources to complete successfully. This change applies automatically to all builds using Elastic build machines. No action is required. Learn more in the build documentation. Read more

DeepSeek V4 Flash Vision Experimental now available on AI Gateway

August 21, 2026

DeepSeek V4 Flash with vision is now available on AI Gateway. This model is an experimental version that accepts images alongside text. You can ask it to describe a picture, read text out of a screenshot, or work through a chart in the same request as your prompt. DeepSeek V4 Flash Vision Experimental now available on AI Gateway. Tool use, reasoning, and caching all work the same as before. Use deepseek/deepseek-v4-flash-vision-exp to get started: Images can be JPEG, PNG, GIF, or WebP. The format is read from the file's own bytes rather than its name or the mediaType you declare, so a mislabeled file still goes through. The -exp in the model ID marks this as an experimental release. Expect behavior to change, and keep a fallback model configured if it's on a production path. To use it in a coding agent, run vercel ai-gateway coding-agents setup to connect agents like Claude Code, Codex, OpenCode, Cursor, and Pi, then select deepseek/deepseek-v4-flash-vision-exp inside the agent. Try Deepseek V4 Flash Vision Experimental in the model playground. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including BYOK requests. View all lang

Always-on tracing for production and preview traffic

August 21, 2026

With always-on tracing, you can now debug your real user requests without reproducing them. Always-on tracing continuously collects traces from your production and preview traffic. Unlike session tracing, which only captures requests from your own browser, it samples your live traffic. You control what's collected with sampling rules. Each rule sets a trace rate for an environment (All, Production, or Preview), optionally scoped to a path prefix like /checkout. Nothing is collected until you add a rule, so you pay only for what you choose to trace. You can also manage sampling rules from the CLI. vercel traces config set adds a rule, or changes the rate of an existing one when you run it again with the same environment and path prefix: Infrastructure and outbound fetch spans are captured automatically. To add framework and custom spans, instrument your app with @vercel/otel. To view a trace, open the Logs or run vercel traces get . Always-on tracing is is now in beta and available to teams on all plans. Tracing is priced at $0.50 per 1M span units. Traces follow your runtime log retention: 1 hour on Hobby, 1 day on Pro and 3 days on Enterprise. Get started by adding a rule in Setti

Microfrontends previews now link across repositories

August 21, 2026

Microfrontends can now link Preview Deployments across repositories when their Git branch names match exactly. In a monorepo, projects share the same Git repository, commits, and branches, so Vercel can automatically link their previews. Cross-repository projects have separate commit histories and branches in separate repositories. With this update, Vercel matches Git-connected projects by branch name and links their corresponding previews. For example, when you preview the new-checkout branch of one project, requests to other projects in the Microfrontends group will use their new-checkout previews when available. This makes it easier to review coordinated changes across repositories in one preview experience. This feature is enabled by default for new Microfrontends groups. You can manage it in your Microfrontends settings. Learn more about how deployment routing works for Microfrontends. Read more

How Ora benchmarks every major AI agent on Vercel

August 21, 2026

Ora on Vercel Front end, back end, and agent runtime on one platform Every major agent tested side by side on live sites Hundreds of commits a day from a 16-person engineering team Ora sends agents onto live websites with instructions to sign up for a product, integrate with it, and pay for it. Agents often fail, and by Ora's estimate, 99% of the web isn't agent-ready. The platform shows customers where and why agents fail, and what to change. Assaf Elovic, co-founder of Ora, spent years helping agents discover the web. His previous company, Tavily, built a web search engine for AI agents and was acquired by Nebius earlier this year. Search solved half the problem, but an agent that finds a product still has to actually use it. He and co-founder Liad Yosef started Ora to measure how ready the web is for agents, and to fix the parts that aren't. Today that means spawning agents against live customer sites from journey.ora.ai, where Ora runs a journey and records the cost, latency, and steps an agent needs to finish a task. The platform runs on Vercel, including the agent runtime. Benchmarking every major agent side by side Every harness expects its own infrastructure Agents decompos

Deployment Storage keeps your deployments rollback-ready

August 21, 2026

Every deployment produces a set of files, including the pages, functions, and assets Vercel serves. Deployment Storage keeps those files available so you can inspect previous deployments and roll back when needed. Instantly roll back to previous deployments in seconds If a production deploy ships a bug or a change you want to reverse, rolling back restores the previous version in seconds. Open your project, click Instant Rollback on the Production Deployment tile, then choose an earlier production deployment. Vercel immediately repoints your domains to that deployment, with no rebuild required. Deployment Storage pricing Deployment Storage is billed at $0.10 per GB per month Hobby teams include up to 10 GB Existing teams continue with their current pricing, with no change to their bill at this time. Control how much Deployment Storage you keep: Edit your Deployment Retention Policy to set how long Pre-Production, Production, Canceled, and Errored deployments stay available. Shorter retention means less storage, but you can't roll back to a deleted deployment The Usage page shows Deployment Storage and Functions Storage by project Keep future deployments smaller by trimming your out

GPT-5.6 Sol is now 50% off a lower price

August 21, 2026

OpenAI lowered list pricing for GPT-5.6 Sol, and the 50% AI Gateway discount now applies to the new, lower price through September 18. Input drops 20%, output drops a third. The discount applies on every OpenAI service tier: Service tier You pay now (input / output) New list price (input / output) You paid before (input / output) Default $2.00 / $10.00 $4.00 / $20.00 $2.50 / $15.00 Flex $1.00 / $5.00 $2.00 / $10.00 $1.25 / $7.50 Priority (fast mode) $4.00 / $20.00 $8.00 / $40.00 $5.00 / $30.00 Rates are per million tokens for requests up to 272K tokens. Cached tokens, cache writes, long-context requests above 272K, and the US regional rates all move by the same proportion. See the pricing page for all model rates. The model ID is unchanged, so requests you already send bill at the new price automatically. BYOK requests bill at your own rate with OpenAI. To use Sol in a coding agent, run vercel ai-gateway coding-agents setup to connect your agents to AI Gateway, then select openai/gpt-5.6-sol in the agent's model settings. See the coding agents guide for setup details. Try it in the playground, or browse every language model on the AI Gateway. Read more

Manage Vercel Container Registry with Vercel CLI

August 20, 2026

You can now manage Vercel Container Registry with Vercel CLI. The new vercel vcr command group lets you: Connect your existing container tooling Build and push images in one step Create and inspect repositories, images, and tags Authenticate your container tooling, such as Docker, Podman, or Buildah, with a short-lived token scoped to your project. Then build and push: The credentials go to Docker itself, so your existing tooling can push to VCR with standard commands: Repositories are created automatically the first time you push, or explicitly with vercel vcr add. From there, you can list what's stored and inspect individual images and tags: Update to the latest Vercel CLI with pnpm i -g vercel@latest and read the Container Registry docs to get started. Read more

Bun 1.4 is now available in Vercel Functions

August 20, 2026

Vercel Functions now support the new Bun 1.4 release, a rewrite of Bun from Zig to Rust, with over 2,900 issues resolved, more than 1,500 additional Node.js compatibility tests passing, and new features and performance improvements. Bun on Vercel Functions: Runs on Fluid compute with Active CPU pricing Supports WebSockets and Bun.serve() Supports common frameworks such as Next.js, Elysia, and Hono To update from the previous Bun version, set bunVersion to 1.4.x in vercel.json. Upgrading is an explicit opt-in because Bun 1.4 has several breaking changes. Review the breaking changes and update your application before upgrading. Learn more about using Bun on Vercel in the runtime documentation. Read more

Manage Vercel Toolbar comments from the CLI

August 20, 2026

You can now use the Vercel CLI to manage Vercel Toolbar comments. vercel comments covers the full triage loop: List and filter unresolved comments for the linked project, scoped to your current Git branch Inspect a full thread with every message and its IDs Reply, resolve, reopen, edit, or delete from the terminal --format json output for scripts and agents Common workflow Start by listing all of the unresolved comments in the current branch: Dig deeper into a specific thread: Reply, resolve, edit: Get started Update to the latest Vercel CLI version and run vercel comments. Or just prompt your coding agent: Learn more in the comments CLI docs. Read more

Custom metrics are now supported in Vercel Observability

August 20, 2026

You can now emit your own metrics directly from your Vercel Functions and analyze them alongside Vercel's built-in observability data. Use the metric() function from the @vercel/functions package to record any value you care about, like request latency or business events, and attach attributes to filter and group by: Once emitted, your custom metrics are available across Observability: • Query builder: slice and aggregate metrics on the fly. • Notebooks: build and share dashboards. • CLI: query using the vc metrics command. Vercel automatically adds useful attributes like the source deployment and function region. Custom metrics are billed as observability events and available on Pro and Enterprise with Observability Plus. Learn more in the docs. Read more

How v0 authenticates to Snowflake without exposing the user's OAuth token

August 20, 2026

AI-generated applications often need to authenticate to external services on behalf of their users. That creates a problem: generated code shouldn't have access to the user's credentials. We faced that decision when building the v0 Snowflake integration. It lets users connect Snowflake, inspect schemas, query data, and generate applications that run against their warehouses. That generated code has to authenticate to Snowflake, but it is written by a model and runs without human review, and prompt injection can steer it into exfiltrating whatever it can read, so the user's OAuth token should never enter the environment the code runs in. We solved this with a Snowflake request proxy for v0 sandboxes, built on the Vercel Sandbox firewall. The sandbox can run normal Snowflake clients, but the real credential is resolved at request time in a server proxy outside the sandbox runtime. This lets existing Snowflake clients work inside the sandbox without exposing the user's credential to generated code. The harder problem was deciding where the proxy could safely inject the credential. The obvious implementation, replacing a placeholder token wherever it appears, introduces another credent

Vercel Agent is now available in Slack code channels

August 20, 2026

Vercel Agent now works in Slack code channels, a new kind of channel launched today for working with a coding agent. Anyone in the channel can follow the work, give Agent new instructions, and review the code it writes. Choose Create a code channel from the Slack sidebar, select Vercel as the agent, and describe what you need to get started. You can also ask Agent to create one from an existing channel or thread. It brings that context with it to the new channel. Use code channels to investigate incidents, write and review code as a team, or manage ongoing work such as dependency upgrades and migrations. Inside a Vercel Agent code channel: Follow the work: Agent posts relevant deployments, logs, errors, and diffs in the channel as it works. Talk without repeated @Vercel mentions: Agent follows the conversation, recognizes which messages are for it, and continues when the team gives it more work. Review pull requests: Ask Agent for a PR review and it posts its findings in the channel. It can find bugs outside the diff, and explain failing checks and their root causes. Approve every change: Agent is read-only by default and never exceeds the requester's permissions. Before making a c

Vercel Connect now supports Microsoft

August 19, 2026

Vercel Connect now includes a managed connector for Microsoft, so your apps and agents can access all Microsoft products such as Teams, OneDrive and more. As a Vercel Managed Connector, Vercel registers the Entra application and configures federated credentials for it, so you don't have to create an app registration or manage client secrets yourself. Authorize the app for your Microsoft tenant once, then attach the connector to the projects and environments it needs. Create a connector from the dashboard or the Vercel CLI: With the connector attached, your code requests a token only when it needs one. The @vercel/connect SDK authenticates with your deployment's OIDC identity and returns a short-lived token for the app itself, using the application permissions your admin granted: Each token is scoped to what you request and refreshed automatically, so there's nothing to leak or rotate. To act as a specific person instead, pass that user as the subject and the token carries their delegated permissions: Read the Vercel Connect docs and create a Microsoft connector to get started. Read more

Chat SDK adds Instagram adapter

August 19, 2026

You can now build bots for Instagram with the new Instagram adapter for Chat SDK. Bots can send and receive DMs and media, render cards as quick replies and link buttons, show typing indicators, receive reactions, and handle story replies. The adapter connects through Meta's Instagram Messaging API and requires a professional Business or Creator account. Messages are buffered, so streamed responses send as one message when the stream completes. Meta enforces a 24-hour messaging window, so bots can only reply within a day of the user's last message. Read the documentation to get started or browse the adapter directory. Read more

Exa joins the Vercel AI Gateway and Agent Marketplace

August 19, 2026

Exa is now available on the Vercel AI Gateway and Agent Marketplace as a native integration. Exa's neural search engine grounds models in current information. Exa search is now a built-in tool on AI Gateway, which runs the search for you without an Exa account of your own. You can also add Exa to a Vercel project from the Agent Marketplace to power search or a research agent, with a single API key that works across every Exa product and billing through your Vercel account. Exa search on AI Gateway Every model on AI Gateway can search with Exa through a built-in tool, with no Exa account or key of your own required. Pass gateway.tools.exaSearch() alongside your prompt and the gateway runs the search, hands the results back to the model, and keeps going until the model stops searching. You get one finished response with a synthesized answer rather than raw results to stream back yourself. The tool takes the same controls as the Exa API, so you can set how many results come back, focus a search on a category such as news or research papers, restrict it to a country, and include or exclude domains. Searches bill through AI Gateway next to your model usage at list price with no markup.

Algolia joins the Vercel Marketplace

August 19, 2026

You can now provision and manage Algolia directly from the Vercel Marketplace. Algolia is a hosted search platform. You send it your content, it builds an index, and your app queries that index instead of your database. Results come back in milliseconds, with typo tolerance, synonyms, and facet counts built in. Key capabilities: Index product catalogs, documentation, or any other content Add recommendations and AI ranking to surface the most relevant results Point the Algolia Crawler at a Vercel domain to index your site automatically Build search interfaces with Algolia's InstantSearch libraries Installing adds your Algolia credentials to your project's environment variables; redeploy to pick them up. Algolia's agent skills install too, so your coding agent knows how to use the API out of the box. Install Algolia on the Vercel Marketplace or with the Vercel CLI: vc i algolia. Read more

Introducing Vercel for Slack

August 19, 2026

Vercel Agent now works in Slack. Mention it the way you'd pull a teammate into a thread, and it reads the discussion, answers with the context of the platform running your app, and turns the team's decisions into changes you approve. Vercel for Slack is available today in Public Beta for Pro and Enterprise teams. Work starts in conversation. An alert gets noticed in a channel, a fix gets agreed on in a thread, and then someone leaves Slack to do the work. With Vercel Agent, you can tag it in for answers about your infrastructure or hand it the work, all without leaving the conversation. What is Vercel Agent? When we introduced Vercel Agent, we called it an AI teammate for your development workflow. Today it's the first responder for your production apps. When something changes, it investigates your logs, metrics, and deployments, finds the root cause, and proposes a fix, often before anyone's opened a laptop. Give it a PR and it flags the regressions and risky changes that a passing CI run won't show. In Slack, that agent joins the conversation itself. It is a core part of Vercel, so your deployments, build statuses, logs, metrics, code reviews, and PRs are in front of it before yo

Vercel Python Queues SDK is now available in beta

August 19, 2026

The Python SDK for Vercel Queues is now available in beta, bringing first-class Python support for running background workloads on Vercel. You publish messages to topics, and independent consumer groups process them in parallel with automatic retries, sharding, and delivery guarantees. Messages can be enqueued or consumed from either JavaScript or Python, meaning a Next.js producer can fan out to both Next.js and Python queue consumers. Publish a message from your Python API: Publish a message from a frontend route handler: Create a subscriber: Opt project into receiving queue message deliveries: Configure cross-runtime consumers: The following configuration allows you to compose a frontend and a backend within a single Vercel project, allowing you to enqueue messages from one service and consume them from another or fan out to multiple. Get started with pip install vercel or visit the Python SDK reference documentation. Read more

Compliance documents are now available in Team settings

August 19, 2026

A new Compliance section in Team settings lets you preview and download Vercel's compliance documents directly in the dashboard. These include attestations like SOC 2 Type 2, security policies, and other security documents that are available through the Trust Center. You can select a single document, an entire attestation, or multiple documents, and download them together as a single zip archive. Every download and preview is watermarked with your user and team information. Some entries link out to external resources instead, such as an auditor's public certificate directory. The Trust Center remains available for additional security information. Vercel's compliance documents are available on Pro and Enterprise plans. Learn more in the compliance documentation. Read more

Fish Audio models now available on Vercel AI Gateway for free

August 19, 2026

Fish Audio's audio models are now available on AI Gateway. To celebrate the launch, every Fish Audio model is free on AI Gateway for the next 30 days, through September 18. Capability Regular Through September 18 Text-to-speech $15.00 per million characters Free Speech-to-text $0.36 per hour of audio Free Four models from Fish Audio are available, including their latest text-to-speech model: fish-audio/s2.1-pro (text-to-speech): Built for low-latency streaming; clones a voice from a reference recording. fish-audio/transcribe-1 (transcription): Returns the text along with the duration of the audio and timestamped segments, down to individual words. fish-audio/s2-pro (text-to-speech): Covers around eighty languages and takes inline tags, plain-language directions written into the text itself, so you can change how a single word or phrase is delivered instead of setting one style for the whole request. fish-audio/s1 (text-to-speech): Reads text that can carry markers for emotion, tone, and sound effects. How to use models during the offer period Using the standard model name (i.e., fish-audio/s2.1-pro) is free, but will automatically begin billing when the offer period ends. To ensure

Vercel for Slack now in public beta

August 19, 2026

Vercel Agent is now available in Slack. Mention @Vercel in any channel or thread, and Agent joins the conversation with relevant context from your Vercel projects, including deployments, builds, logs, metrics, configuration, and connected repos. For example, use it to: Investigate incidents: Ask why a deployment failed or what’s causing a runtime error. It investigates and explains what it found. Write and review code: Work with your team and Agent to fix build and CI failures, review pull requests with production context, or turn a decision from the thread into a tested PR. Take action with approval: Tell it to make a change, such as rolling back a deployment or updating configuration. Agent drafts a plan and waits for your approval. Once approved, it gets temporary access to carry out the plan, then returns to read-only. Vercel Agent acts through its own identity. It's read-only by default, never exceeds the requester’s permissions, and makes no changes without approval. Every action is attributed to Agent, the requester, and the approver. Vercel for Slack follows Vercel Agent pricing. During the beta, a limited number of simple requests are free. Paid work is billed on demand. T

$1 million hacker challenge for Vercel Sandbox

August 18, 2026

Agents need to run untrusted code, and the microVM has become the standard way to do it: a dedicated guest kernel per workload, isolated from the host and from every other workload on the same machine. But recent security research and real-world incidents have revealed that agents running untrusted code do not need to cross a VM boundary to escape containment; they only need one network path the security model failed to account for, which we explained in our post A sandbox without a network boundary is only half a sandbox. Isolation only holds if both sides of it hold: the Firecracker microVM and the host-side network controls. Recently our CTO pointed an open-weight model with no safeguards at Vercel Sandbox. It did not escape, but it mapped the guest kernel, built a VM to reproduce its ideas, and wrote a fuzzer. Defenders have first-mover advantage, but it won't last forever, and the choice is when to test the boundaries (we strongly encourage building a scanning program, which you can do on any budget with an open-source tool like deepsec and AI Gateway). We are proactively choosing to test Vercel Sandbox on our own schedule, not an attacker's, and we are doing it in the open, w

$1 million hacker challenge for Vercel Sandbox

August 18, 2026

Agents need to run untrusted code, and the microVM has become the standard way to do it: a dedicated guest kernel per workload, isolated from the host and from every other workload on the same machine. But recent security research and real-world incidents have revealed that agents running untrusted code do not need to cross a VM boundary to escape containment; they only need one network path the security model failed to account for, which we explained in our post A sandbox without a network boundary is only half a sandbox. Isolation only holds if both sides of it hold: the Firecracker microVM and the host-side network controls. Recently our CTO pointed an open-weight model with no safeguards at Vercel Sandbox. It did not escape, but it mapped the guest kernel, built a VM to reproduce its ideas, and wrote a fuzzer. Defenders have first-mover advantage, but it won't last forever, and the choice is when to test the boundaries (we strongly encourage building a scanning program, which you can do on any budget with an open-source tool like deepsec and AI Gateway). We are proactively choosing to test Vercel Sandbox on our own schedule, not an attacker's, and we are doing it in the open, w

Sign JWTs from your Functions without managing private keys

August 18, 2026

Vercel KMS lets you sign JWTs and arbitrary messages from your Vercel Functions using managed asymmetric signing keys, so private keys never live in your code or environment variables. Your function authenticates with its Vercel OIDC token, and the private key stays inside Vercel's key management service while verifiers use only the public key. With Vercel KMS you can: Create and rotate issuers and signing keys (RSA, ECDSA, and EdDSA) from the CLI and dashboard. Sign JWTs with custom claims and a configurable TTL, or sign raw bytes, using the @vercel/kms package. Grant a project signing access per environment, including production, preview, development, and custom environments. Constrain the claims a project can request per grant, and validate token claims against a JSON Schema. Verify signed tokens anywhere. Each issuer publishes a public OpenID Connect Discovery document at https://kms.vercel.com//.well-known/openid-configuration and a JWK set at https://kms.vercel.com//jwks.json, so any standard OIDC or JOSE library can validate tokens without Vercel-specific code. Sign a token inside a function: Verify a token anywhere with a standard JOSE library, using the issuer's public JWK

Vercel for Platforms can now deploy from your users' GitHub repositories

August 18, 2026

Teams building on Vercel for Platforms can now create deployments directly from their users' GitHub repositories, without requiring them to install the Vercel GitHub App. When creating a deployment, pass a gitAccessToken alongside gitSource. Vercel uses the token to retrieve the source and build the deployment. Tokens should be read-only, scoped to the requested repository, and valid for 24 hours or less. Vercel temporarily stores the token in encrypted form so source retrieval can complete. The token is never stored on the deployment. Learn more about using a GitHub access token. Read more

The Vercel Sandbox CLI is now more agent-friendly

August 18, 2026

The Sandbox CLI version 4.0.1 ships four output and error improvements: New sandboxes point users towards how to connect into them. A sandbox not found error shows where to look for existing sandboxes. Common mistakes, like typing a command after sh, redirect to the command that was intended. Outdated installs show an update notice when they create a new sandbox. Each of these messages now carries the action to take next instead of leaving you to work it out. New sandboxes tell you how to connect create and fork output now ends with a connect hint, so a fresh sandbox never leaves you wondering what to type next. Not found errors point you to your sandboxes When a command references a sandbox that does not exist, the error is quiet and actionable instead of a raw API dump. Setting DEBUG=sandbox:errors restores the request detail for debugging. A misused command explains the right one sh opens a plain interactive shell, so it takes no command of its own. The CLI can now return precise answers that rewrite the attempt into the command that does what you meant, whatever the command is. Outdated installs warn you at creation time create, sh, and run check for a newer published CLI and p

Cline is now available in the AI SDK harness layer

August 18, 2026

The AI SDK harness layer lets you run established coding-agent runtimes through one unified interface, so you can switch runtimes without changing your application code. Today we are adding Cline, which runs through the same HarnessAgent interface as other supported harness. @ai-sdk/harness-cline is the official harness adapter for Cline, implemented in collaboration with the Cline team. Cline runs fully in the host process and uses the sandbox as a remote filesystem and shell, so only its tools operate in the sandbox. No bridge process is installed inside the sandbox, similar to the existing Pi harness. Read the Cline harness documentation to get started. The full supported list of harnesses is now: Claude Code, Cline, Codex, Deep Agents, Grok Build, OpenCode, and Pi, with more coming soon. Read more

GLM 5.3 now available on AI Gateway

August 18, 2026

GLM 5.3 from Z.ai is now available on AI Gateway. GLM 5.3 has improvements vs. GLM 5.2 at complex software engineering and at agent tasks that run across many steps, and it reaches those results while producing fewer output tokens than GLM 5.2 did at the same effort level. Z.ai also reports stronger vulnerability discovery, including reasoning across the successive stages of an exploitation chain. DeepsecBench measures that work directly, scoring how reliably a model finds security vulnerabilities in application code and what a run costs, and it covers GLM-5.3 at both high and max effort. GLM 5.3 takes text input, with a 1M token context window and a maximum output of 128K tokens, both unchanged from GLM 5.2. It supports function calling, structured output, streaming, and context caching. To use the model, the slug is zai/glm-5.3: To use it in a coding agent, run vercel ai-gateway coding-agents setup to connect agents like Claude Code, Codex, OpenCode, Cursor, Pi, and more, then select zai/glm-5.3 inside the agent. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider

Deploy Cursor Origin repositories with Vercel in public beta

August 17, 2026

You can now connect Origin repositories to Vercel. Once connected, pull requests on Origin will automatically generate Preview deployments on Vercel, and merging will trigger a Production deployment. You can connect Origin through Team settings, Project settings, or by clicking "Continue with Origin" when creating a new project. You can also connect Vercel to a repository from within Origin by adding Vercel in the repo's Apps tab. Vercel's integration is in public beta and Origin is in early beta. Origin repositories are subject to Vercel's existing private repository policy. Learn more about connecting Origin in the Git integration documentation. Read more

GPT-5.6 Sol is 50% off on AI Gateway for the next month

August 17, 2026

GPT-5.6 Sol, the flagship of OpenAI's GPT-5.6 series, is 50% off on AI Gateway through September 18. The discount applies on the OpenAI provider to all token types, tiers, regions, and modes, and it is available only on requests running directly through AI Gateway (not BYOK). Pricing: 50% off Service tier New price per M tokens (input / output) Original price per M tokens (input / output) Default $2.50 / $15.00 $5.00 / $30.00 Flex $1.25 / $7.50 $2.50 / $15.00 Priority (fast mode) $5.00 / $30.00 $10.00 / $60.00 The discount applies on the same terms everywhere else the model is priced: across every service tier, including fast mode, and to cached tokens, cache writes, long-context requests, and different regions. It covers requests billed through AI Gateway on the OpenAI provider. BYOK requests run on your own provider accounts and bill at whatever rate you have with them. The model ID is unchanged, so requests you already send pick up the discounted rate with no code change: Sol takes a reasoning effort up to max for the hardest problems, accepts text, image, and PDF input, and carries a long context window. To use it in a coding agent, run vercel ai-gateway coding-agents setup to

Encrypted Client Hello (ECH) is now supported on Vercel CDN

August 14, 2026

Vercel CDN now supports Encrypted Client Hello (ECH) for domains managed by Vercel DNS. ECH encrypts the Server Name Indication (SNI) in the TLS handshake, the last part of an HTTPS connection that revealed which hostname a client was connecting to. With ECH, network observers see a connection to Vercel's shared ECH hostname, vercel-ech.com. ECH is managed at the platform level and automatically enabled selectively where supported, starting in the US for a limited number of TLDs, and expanding to more TLDs and regions over time. Recent versions of Chrome, Edge, and Firefox support Encrypted Client Hello. Learn more about ECH and how Vercel CDN protects your application in the CDN encryption documentation. Read more

Gemini 3.7 Flash now available on AI Gateway for 50% off

August 13, 2026

Gemini 3.7 Flash from Google is now available on AI Gateway for 50% off till December 31st, 2026. Gemini 3.7 Flash improves on prior Flash models at software engineering and agentic work. It resolves issues more reliably and spends less time stuck in failed agent loops, which matters on long tool-calling sequences where one derailment costs the rest of the run. It also generates desktop and web application code directly from design mocks, with closer adherence to the source design. To use Gemini 3.7 Flash, set model to google/gemini-3.7-flash in the AI SDK: To use it in a coding agent, run vercel ai-gateway coding-agents setup to connect Claude Code, Codex, OpenCode, or Pi, then select google/gemini-3.7-flash inside the agent. To try Gemini 3.7 Flash with no code, try the model in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on infe

One-click upgrade for deprecated Node.js versions

August 13, 2026

Team owners and members can now upgrade all projects using Node.js 20 or earlier to Node.js 24 with one click from the Vercel Dashboard. This updates the Node.js version in project settings. If your project also defines a version in package.json, you'll need to update it manually. Existing deployments are not affected. View and upgrade deprecated Node.js projects now. Read more

Grok Build is now available in the AI SDK harness layer

August 13, 2026

The AI SDK harness layer lets you run established coding-agent runtimes through one unified interface, so you can switch runtimes without changing your application code. Today we are adding Grok Build, which runs through the same HarnessAgent interface as every other supported harness. @ai-sdk/harness-grok-build is the official harness adapter for Grok Build, built on top of the ACP harness adapter (@ai-sdk/harness-acp). Read the Grok Build harness documentation to get started. The full supported list of harnesses is now: Claude Code, Codex, Deep Agents, Grok Build, OpenCode, Pi, with more coming soon. Read more

Use ACP-compatible harnesses with the AI SDK harness layer

August 13, 2026

The AI SDK harness layer now supports any Agent Client Protocol (ACP)-compatible harness with HarnessAgent through the new @ai-sdk/harness-acp package. Previously, every harness adapter wrapped one specific runtime (Claude Code, Codex, Pi, Deep Agents, OpenCode). @ai-sdk/harness-acp wraps the protocol instead. It is a meta adapter: rather than adapting a single harness, it lets you build an adapter for any harness that ships an ACP-compatible package. Implement a harness by passing that package to createACP and configure basic harness mapping. Then pass it to HarnessAgent like any other harness: ACP is an abstraction over coding harnesses, but the AI SDK harness layer stays deliberately decoupled from it. Not every harness supports ACP, and for some, ACP limits or changes how well they expose their internal behaviors, so a direct adapter can produce a tighter integration. For Claude Code and Codex, prefer the direct @ai-sdk/harness-claude-code and @ai-sdk/harness-codex adapters over an ACP-based implementation; reach for @ai-sdk/harness-acp when a harness has no direct adapter but ships an ACP-compatible package. Read the ACP harness documentation to get started. Read more

Inside the Vercel intern experience

August 13, 2026

The best way to learn is by shipping product to users in production, and that's why we don’t treat our interns as temporary guests. Every intern at Vercel is expected to operate like a core part of our engineering team from day one. This past winter, we welcomed a small cohort of those interns to our headquarters in San Francisco. Over an intense four months, they contributed to some of our most important products: the CDN, v0, financial infrastructure, AI Gateway, and more. They shipped features used by millions, helped close the gap for enterprise deals, and in more than a few cases, left a mark that will outlast their time here. Meet the Winter '26 intern cohort Yash Kothari, CDN Team University of Waterloo What Yash shipped: Yash's internship spanned the full CDN stack, from infrastructure to CLI to UI in the dashboard. His first major project was Project-level routing, which is part of the new CDN tab in the Vercel dashboard. This feature lets customers update routing rules, including redirects, rewrites, and header transforms, from the dashboard, API, CLI, or SDK, without needing to redeploy their project or touch a single line of code. Yash built the underlying API endpoints

GLM 5.2 free for eve agents through August 27 via Blackbox on AI Gateway

August 13, 2026

GLM 5.2, the open-weights coding model from Z.ai with a 1M-token context window, is free for eve agents through August 27, served by Blackbox AI on AI Gateway. New eve agents come with GLM 5.2 as their default model. Use npx eve@latest init my-agent to get started. Existing agents can benefit from this as well with a one-line change in agent/agent.ts: Or run eve set from the CLI: After August 27, GLM 5.2 stays available on AI Gateway at standard provider rates. Try GLM 5.2 in the model playground. Read more

Set up coding agents in one command with AI Gateway

August 12, 2026

Using coding agents means setting up multiple accounts, provisioning API keys, and scattering observability and billing. Now, you can route them through AI Gateway to centralize all of this and add controls, with set up in one command: Any of 200+ models in any agent, including models the agent has no native support for. The gateway routes the request to whichever model you picked, with automatic fallbacks. One dashboard for agent traffic: spend, performance, tokens and more, across the models your agents used. Budgets, resets, and expiry on the keys agents use to control spend. Team-wide policy your agents can't route around: turn on Zero Data Retention or restrict which providers may serve requests, and it holds for every agent request without editing an agent's config. Request traces, with the cost, tokens, and model behind each request. To start: vercel ai-gateway coding-agents setup connects Claude Code, Codex, OpenCode, and more in one command. It finds the agents on your machine, creates the API key, and writes the gateway URL and your credentials into each agent's own config format. Agent Selecting a model Claude Code, CLI and desktop /model or /model , as in /model anthrop

Free domain for one year, now for all Pro teams

August 12, 2026

We recently announced that new Pro teams are eligible to claim a free domain for their first year. We are now extending this offer to all Pro teams. Existing Pro teams can now also claim a free domain for one year. The offer covers one eligible domain per team across six TLDs: .online, .site, .space, .store, .tech, and .website. After the first year, your domain will renew automatically at the regular price shown at checkout when you claim it. Claim yours at vercel.com/domains, and see the free domain FAQ for full terms. Read more

Grok 4.6 now available on AI Gateway

August 12, 2026

Grok 4.6 from SpaceXAI is now available on AI Gateway. The model has a 500K token context window and accepts text and image inputs. Grok 4.6 supports low, medium, high, and xhigh reasoning levels and defaults to high. To use Grok 4.6, set model to xai/grok-4.6 in the AI SDK: To use it in a coding agent, run vercel ai-gateway coding-agents setup to connect Claude Code, Codex, OpenCode, or Pi, then select xai/grok-4.6 inside the agent. To try Grok 4.5 with no code, try the model in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Read more

DeepSeek V4 Pro now runs updated weights on AI Gateway

August 12, 2026

DeepSeek V4 Pro now runs on updated weights on AI Gateway. They are used by default when you call deepseek/deepseek-v4-pro, so existing requests pick them up with no change to the model ID or your code. To use the updated DeepSeek V4 Pro, set model to deepseek/deepseek-v4-pro-0813 in the AI SDK. AI Gateway will route to providers with the new weights: To run V4 Pro in a coding agent, use vercel ai-gateway coding-agents setup to connect your agents to AI Gateway, then select deepseek/deepseek-v4-pro-0813 in the agent's model configuration. See the coding agents guide. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Exa web search free through August 31 on AI Gateway and eve

August 12, 2026

Exa web search is now free on AI Gateway through August 31, and it's now the default web search for eve agents. The tool works with any AI Gateway model, with no separate Exa API key. When the model calls it, AI Gateway routes the request to Exa's Search API, returning web results and extracted page content with support for domain filters, date filters, and token-efficient excerpts. The built-in web_search tool in eve now uses Exa by default when the model is served through AI Gateway. See other supported search configurations in the eve docs. To use it with the AI SDK, pass gateway.tools.exaSearch() to the tools parameter of a generateText call: Read the web search documentation to learn more and get started. Read more

Building a software factory for AI SDK

August 12, 2026

The AI SDK is one of the most popular open-source AI projects in the world. It serves over 20 million npm downloads a week and the repo has over 26,000 stars. Maintaining the codebase means tracking four moving targets at once: Model providers: new providers, new capabilities, and new bugs UI frameworks: bindings for React, Next.js, Svelte, Vue, and others Sandboxes: the execution environments agents run code in Harnesses: adapters for Codex, Claude Code, Pi, and others After multiple years of growth, the repo was getting 100+ new issues every month, and when Anthropic's Opus 4.6 model was released, PRs hit an inflection point. By late June, that compounding had accumulated over 1,000 open issues and almost 800 pull requests. That backlog is not a discipline problem. No maintainer, however good, can close that gap by working harder, and because generating code is cheap, it will only grow. Instead of trying to scale ourselves, we built a software factory. Four weeks in, it authors between 25 and 35% of PRs we merge and closes 70-80% of issues. Deciding what type of factory to build Before we built anything, we had to answer three questions: Why our existing approach using agents was

LaunchDarkly is now available on the Vercel Marketplace

August 11, 2026

LaunchDarkly is now available on the Vercel Marketplace, allowing you to quickly get started with feature flags without additional setup. You can: Sync flags into Global Config and evaluate them locally Target releases by user, attribute, or segment Run experiments with metrics, and roll back with a kill switch View and override flags from the Vercel Toolbar with the Flags Explorer To get started, run vercel install launchdarkly, add the @flags-sdk/launchdarkly adapter, and declare a flag with the Flags SDK: Using a coding agent? Hand it this prompt: Add LaunchDarkly from the Vercel Marketplace, or read the adapter docs. Read more

Vercel Connect now supports CLI setup for 100+ connectors

August 11, 2026

You can now integrate 100+ services through Vercel Connect from the CLI. Previously, vercel connect create completed setup in the terminal for some services, and opened the dashboard for everything else. Pass the service name to create a connector, then attach it to your project: The CLI pre-populates the brand name, icon, auth type, and MCP or discovery URL, then prompts for any credentials the service requires and creates the connector. After attaching the connector, call getToken with its UID (service/name) to request a short-lived, scoped token from the provider: Get started by browsing the connectors directory, or read the CLI reference. Read more

Everything hackable will get hacked

August 11, 2026

Over the past year, AI models have become much more capable of performing cybersecurity work. These changes are reshaping both the threats facing the web and the tools available to defend it. Right now, defenders have an advantage because they can use stronger models for defensive work than the open-weight models broadly available for offensive research. But this advantage will not always last. The gap will soon close. I have both good news and bad news, neither of which is yet widely understood in the community. Bad news: Near-frontier open-weight models that perform offensive security research are available today. Kimi K3 is an Opus 4.X-class model with no relevant cybersecurity safeguards. Good news: You do not need to wait for “Mythos access” or OpenAI’s cyber program to begin defensive cybersecurity work. Frontier models, with the notable exception of Fable 5, will perform defensive cybersecurity tasks today. The uncertainty around Mythos 5's release seems to have created a kind of paralysis among defenders, many of whom are underutilizing the powerful tools available to them today. This post aims to give readers insight into how we're thinking about AI-enabled defensive cyber

DeepSeek overtakes Google on volume, cost per token falls 13.6%

August 11, 2026

AI Gateway Production Index — August 2026 Every month, AI Gateway routes tens of trillions of tokens between production applications and AI labs. That traffic gives us a view of what AI usage actually looks like in today's enterprise, and we publish it here monthly. See the Production Index reports from May, June, and July. August 2026 Summary The August index reports on AI Gateway data collected through July 2026. The average price paid per token fell 13.6% in July, after rising almost 20% in May and holding steady in June. Token consumption grew so quickly that even with 37% growth in spend, cost per token saw a double-digit drop. DeepSeek became the second-largest lab by token volume, now running more than twice Google's volume. Anthropic collected 65% of gateway spending on 30% of token volume, at 4.4 times the average price of every other lab's tokens. Both media leaderboards changed hands. Google's Nano Banana took the lead in image volume from OpenAI's GPT Image, and ByteDance's Seedance led video in both volume and dollars. Kimi K3 launched into agent work Moonshot released Kimi K3 on July 16. Like Z.ai's GLM 5.2 released in June, it is built for long-horizon agent work, so

A sandbox without a network boundary is only half a sandbox

August 11, 2026

Running untrusted code safely requires more than separating it from the host. You also have to control what that code can reach. This matters more as AI agents gain the ability to read files, execute commands, install packages, and generate programs of their own. A microVM can prevent that code from accessing the host or another workload. By itself, it cannot stop the code from exfiltrating data, probing internal services, attacking systems elsewhere on the internet, or using credentials available inside the environment. Isolation without egress control contains the process, not its consequences. A complete sandbox therefore needs both compute isolation and control over the authority available through its network: where code can connect, which credentials it can use, and how those permissions change throughout the workload’s lifecycle. These controls are part of the security boundary, not protections to bolt on later. A sandbox has more than one boundary Compute isolation answers one important question: what can this program access on the machine where it runs? Network isolation answers another: what can it access, or attack, through the network? Consider an agent that reads a repo

How we migrated the database behind every Vercel build

August 11, 2026

Every build on Vercel starts in the build warm pool, which is a set of standby containers that let builds begin without waiting for new compute. The pool runs on state that tracks which containers are ready, the tokens each one uses to authenticate, and the mapping that ties every running build back to the deployment that gets billed for it. When we built the pool, we put all of that in Redis, which was fast and made sense at the time. Over the years, though, that state turned into a liability. Tokens and container statuses can be rebuilt if they get lost, but the billing mappings can't, and all of it was sitting in a store that we ran as an ephemeral cache. That state needed to live somewhere durable, which is why we decided to migrate it to DynamoDB. The problem is that the pool never stops. Containers are coming up, polling, picking up work, and expiring around the clock, which meant we couldn't pause the world, copy the data, and restart. The migration had to happen live, under production traffic, in phases, each one behind a flag with a rollback ready if we needed it. Durable state in an ephemeral store Redis was a good home for it at first. It was fast, familiar, and efficien

Vercel Connect adds observability support

August 11, 2026

Vercel Connect now gives teams line-level visibility into the token lifecycle. See who created a token, what app or project used it, when, and whether it is still active. Every connector's detail page includes a new Observability tab: Runtime events: Every token request, authorization, refresh, revocation, and trigger delivery, filterable by type. Correlation IDs: Stable tokenId and authorizationId values link each token's events, so you can match Connect events to your own systems. Activity: The Activity filter button opens the Activity page pre-filtered to the connector's configuration events, so you can see when and how it changed. Connect observability is available on all plans. Events are retained for 12 hours on Hobby, 3 days on Pro, and 30 days on Enterprise. To retain events longer, forward them to a custom webhook endpoint by adding a Drain on Pro and Enterprise plans. Enterprise teams also get role-based access control and audit logs. Get started in the Observability tab on a connector, or read the documentation. Read more

Enterprise Managed Users is now generally available

August 11, 2026

Enterprise Managed Users (EMU) gives organizations full control over the Vercel accounts tied to their verified domains. It makes the organization's identity provider the single source of truth for authentication and account lifecycle, so accounts on company domains are governed centrally rather than owned by individuals. Now generally available IdP-enforced sign-in: Managed users sign in through SAML SSO only. All other login methods, including email OTP, GitHub, Google, and GitLab, no longer work for managed accounts. SCIM provisioning & lifecycle: Directory Sync creates, updates, and deprovisions managed users automatically, keeping accounts in sync with your IdP. Owned by your organization: Profile settings are managed through your IdP rather than by the individual. Self-serve enablement: Turn EMU on for a verified domain in Team Settings under Security & Privacy, and extend it across multiple eligible teams in a single step. Requires an Enterprise plan, enforced SAML SSO, active Directory Sync, and at least one verified domain. In Beta, available on request Hobby team transition: Members with an existing personal account on a newly managed domain are guided to add a personal e

Vercel Sandbox now runs on Vercel Managed Images

August 10, 2026

Today we are introducing Vercel Managed Images (VMI), a set of versioned, open-source base images you can use as-is or extend. The source for every image lives in the public vercel/sandbox repository. Managed images replace Sandbox runtimes, which are now deprecated. Starting with version 3 of the Sandbox SDK, new sandboxes default to vercel/sandbox/universal:latest. It ships with Node.js, Python, common coding agents and standard utilities, so most users never build a custom image or install packages at boot. The universal image switches our default operating system from Amazon Linux to Ubuntu, a lighter and more widely used system across the industry. Secure by default Every managed image gets a nightly release. Rolling tags such as latest and the major-version tags pick up operating system and dependency updates automatically, including security patches and new releases of Node.js, Python, and the preinstalled coding agents. Within each release, dependencies are pinned to specific versions wherever possible, so a given image version stays consistent. If you need a fully immutable, reproducible environment, pin to an image digest (SHA). Digest-pinned images opt out of automatic u

Simplified onboarding for deepsec

August 10, 2026

deepsec, the open-source security review harness from Vercel, now lets you set up a repository and run its first security review with a single command. The init command now automates the standard setup process: creates the isolated .deepsec/ workspace, the only thing added to your repository, and installs its dependencies configures model access, through Vercel AI Gateway or your own provider keys generates a description of the codebase and its attack surface that every later review depends on runs a pattern scan, generating extra scan patterns where the built-in set leaves coverage gaps starts the AI review of the flagged files Setup is checkpointed after each step. If a run stops, whether from closing the process, a failed step, or a hit cost or duration limit, re-running init resumes from the last completed step. Run npx deepsec init to start your first scan, or read the docs to learn more. Read more

Bun runtime for Vercel Functions now accepts Bun.serve as an entrypoint

August 10, 2026

The Bun runtime for Vercel Functions now supports Bun.serve() as a function entrypoint, including WebSocket handlers. The server you run locally with Bun deploys as-is, without being wrapped in a framework. Enable the runtime by setting "bunVersion": "1.x" in vercel.json. Deploy a routes-based server Create a server with a routes map in server.ts at the project root. Accept WebSocket connections Add a websocket handler and call server.upgrade(request) in fetch to upgrade matching requests. The rest of the server stays the same. WebSocket connections run on Fluid compute with Active CPU pricing, so you pay only for time spent processing messages, not idle connection time. A connection is pinned to one function instance for its lifetime, and a single instance can handle multiple concurrent connections. Use an external data store to coordinate messages across instances. Read the documentation to get started. Read more

Grok Imagine Image 2.0 now available on Vercel AI Gateway

August 8, 2026

Grok Imagine Image 2.0 Preview from xAI is now available on AI Gateway. The model follows detailed instructions closely and plans typography and layout together, so dense, multi-part visuals like infographics, posters, and title screens hold their structure and small text stays legible. Grok Imagine Image 2.0 Preview also supports image editing, keeping subjects and details consistent across repeated generations. Try the model out now in imagine.vercel.sh, running on AI Gateway. To use Grok Imagine Image 2.0 Preview, set model to xai/grok-imagine-image-2.0-preview and call generateImage from the AI SDK. Set resolution to 1k or 2k under providerOptions.xai to pick an output tier, and n for more than one image per call. For image editing, pass an image in prompt.images alongside the instruction, and the model changes what you asked for and leaves the rest: To view all image models supported on AI Gateway, see the full list here. Read more

Audit Log Drains now support Datadog, Splunk, and Panther

August 7, 2026

Audit Log Drains now stream your team's audit events into Datadog, Splunk, and Panther, joining the existing custom HTTPS endpoint and Amazon S3 destinations. An Audit Log Drain forwards every event from your team's Activity Log, plus additional audit metadata, to the destination you choose. They're available on Enterprise plans. To create one, go to Drains in your team settings. Click Add Drain, choose Audit Log as the data type, and pick a destination. Audit Log Drains replace Custom SIEM Log Streaming. If you already stream audit logs to a SIEM, follow the migration guide to move your integration over. Learn more about Drains in the documentation. Read more

Skill packs are now available on skills.sh

August 7, 2026

You can now bundle multiple agent skills into a shareable pack on skills.sh. Build a pack from community skills, your own local folders and zips, or public and private GitHub repositories. Every pack is unlisted, with its own URL. Share that link with a single person, or your whole team to standardize the skills across your projects. Install a pack with one command: Run npx skills update to update a pack to its latest version. Explore and create packs at skills.sh/packs. Read more

Free domain now included with new Pro subscriptions

August 7, 2026

New Pro subscriptions now include a domain free for the first year. The offer covers six TLDs: .online .site .space .store .tech .website Each team has 30 days after subscribing to claim one domain, either at checkout or later from the domain search in the dashboard. After the first year, your domain will renew automatically at the regular price shown at checkout when you claim it. See the free domain FAQ for the full terms. Read more

Vercel AI Gateway and Vercel Sandbox now available on Hermes Agent

August 7, 2026

Hermes Agent can now use Vercel AI Gateway as its inference layer and run its agent commands in an isolated Vercel Sandbox microVM. Access 200+ models through AI Gateway at no markup on tokens, and every request appears in your AI Gateway dashboard with the rest of your usage and spend. Setup Install Hermes, then pick Vercel AI Gateway in the setup wizard: The picker pulls live model availability and current pricing from AI Gateway. For the full walkthrough, see Configuring Hermes. Running commands in Vercel Sandbox The Vercel Sandbox backend is opt-in. Commands run locally until you set terminal.backend to vercel_sandbox; with it enabled, agent commands run in a cloud microVM with a workspace root of /vercel/sandbox instead of on your machine. The backend supports node24 (default), node22, and python3.13 runtimes. For local development, use VERCEL_OIDC_TOKEN. Run vercel link and vercel env pull to get a development token. To get started, read the Hermes documentation and the Vercel Sandbox documentation, or browse AI Gateway models. Read more

Vercel Container Registry repositories can now be made public

August 7, 2026

Vercel Container Registry now lets you make a repository public, so anyone with a Vercel account can pull and use its images. Sharing already granted read access to up to 100 teams, and making a repository public opens that access to every Vercel team rather than a named set. Public access is read-only, so anyone can pull and use the images but cannot push, delete, or otherwise change the repository. Repositories stay private by default. Make a repository public from the project dashboard under Images → repository → Settings → Public Access, where confirming requires typing the repository name. You can also set visibility from the Vercel CLI: Public images work like shared images in Vercel Sandbox, which accepts a team-scoped image reference in Sandbox.create(): Update your repository visibility from the dashboard, the CLI, or the REST API, and see the Vercel Container Registry documentation to learn more. Read more

Introducing Agent Plugins

August 6, 2026

via Codex

Build a plugin once and use it across compatible agent clients. OpenAI introduced Agent Plugins, an open standard developed with AWS, Cursor, GitHub, VS Code, and Vercel that packages Agent Skills and supports MCP server configurations in a shared format.

View on X

Ling 3.0 Tiny is now available on AI Gateway

August 6, 2026

Ling 3.0 Tiny from ANT Group is now on AI Gateway, free to use till 8:00am PT on 8/14. Ling 3.0 Tiny takes the free slot from Ling 3.0 Flash. Ling 3.0 Tiny is a MOE model with 7.9B total parameters and about 1.3B active per token, a 256K token context window, and up to 32K output tokens. The model is built for responsive agents, instruction following, and multi-turn conversation, with native function calling and prompt caching. To use Ling 3.0 Tiny, set model to inclusionai/ling-3.0-tiny-free in the AI SDK. On August 14th, the new model name will be inclusionai/ling-3.0-tiny. Try Ling 3.0 Tiny in the model playground. To use it in a coding agent, run vercel ai-gateway coding-agents setup and select inclusionai/ling-3.0-tiny-free inside the agent. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. R

Introducing Agent Plugins 1.0.0

August 6, 2026

Agent Plugins 1.0.0 is now available. It is an open, vendor-neutral standard for packaging Agent Skills and MCP servers into portable plugins. Compatible agent clients can discover and load them. Agent Plugins defines a common format: a root plugin.json manifest, plus fixed locations for portable components. Each client keeps control of installation, distribution, policy, user experience, and client-specific capabilities. At launch, Agent Plugins is supported across the following clients: ChatGPT and Codex Cursor GitHub Copilot Kiro VS Code Read the specification and implementation guides, learn more in the announcement post, and follow our guide to build a plugin on the new standard. Read more

Pause workflows for approval with Chat SDK

August 6, 2026

You can build human-in-the-loop approvals with Chat SDK through the new chat/workflow subpath. One requestApproval call posts a card with Approve and Deny buttons and suspends a Workflow SDK workflow until someone decides. The wait can last seconds or days and survives deploys and restarts. You don't need an approvals table, an onAction handler, or a polling loop: Start the workflow from any handler, and Thread instances serialize across the workflow boundary automatically. The card handles the rest: Scoped approvers: Pass approvers to restrict who can decide. Clicks from anyone else post a notice, and the workflow keeps waiting. Verified decisions: Chat SDK checks the platform's signature on every click, so the user.id on the result is the person who decided. Audit trail: Once a decision lands or the timeout elapses, the card is edited in place with an outcome line, preventing stale clicks. Get started by reading the approvals documentation. Read more

Give every agent in Herdr its own Vercel Sandbox

August 6, 2026

Terminal coding agents like Claude Code, Codex, and OpenCode can now each run in their own isolated Vercel Sandbox, orchestrated from Herdr, a tmux-style manager that runs them side by side in panes. Nothing an agent runs or edits touches your machine. When you start an agent, the plugin uploads your project files into a new Sandbox, installs the agent there, and opens a pane connected to it. The pane works like a local terminal; you type to the agent and watch it work. But it lives in the Sandbox, not on your machine, so you can run several agents at once without slowing anything down. Everything the plugin does is machine-readable. Each action returns its results as JSON, so whatever you can do from a pane, like starting an agent, checking its state, or applying its changes, a script or another agent can do too. A few safeguards are built in. The plugin shows every file before the first upload, returns remote changes as a Git patch you apply only when you choose, and requires a human to type DELETE to permanently remove a Sandbox. Claude Code (2.1.220), Codex (0.146.0), and OpenCode (1.18.9) are all verified and available by default. To get started, install the plugin: Run vercel

Introducing Agent Plugins

August 6, 2026

Today, Agent Plugins 1.0.0 is publicly available. Agent Plugins is an open, vendor-neutral standard for plugins that extend AI agents. Agent Skills provide reusable instructions and resources for AI agents. MCP servers connect agents to tools and services. Both can be reused across clients, but clients often package and discover them differently. Agent Plugins gives compatible clients a common format: a directory with a plugin.json manifest and fixed locations for its components. The format is intentionally small and easy to implement, and it leaves installation, distribution, policy, user experience, and client-specific capabilities to each client. One package for the portable parts Extension authors often adapt the same component to several client formats. Even though the underlying Skill or MCP server is identical, clients often expect different top-level metadata, discovery paths, or MCP configuration. Agent Plugins gives those shared components one predictable, structured home: A minimal JSON manifest (plugin.json) identifies the specification version and names the plugin: Those two fields are the minimum requirement for the manifest, and the rest of the contract is represente

Seedance 2.5 now available on Vercel AI Gateway

August 6, 2026

Seedance 2.5 from ByteDance is now available on AI Gateway. It can generate up to 30-second clips with synchronized audio, and takes up to 50 reference inputs in one request, spanning images, video, audio, and style. A single clip holds camera movement and continuity without stitching shots together in post, and short clips can be extended, carrying over character, scene, and camera movement. Prompts work in more than ten languages. Seedance 2.5 can also edit a finished video in place, swapping backgrounds, products, or characters while the frame, camera, and pacing stay fixed, so one shoot yields several variants. Generating a video Set model to bytedance/seedance-2.5 and call generateVideo from the AI SDK: Adding image references Pass image URLs in inputReferences and point at them in the prompt with [Image 1], [Image 2]. Tag each entry with its media type. Seedance treats an untagged URL as an image. Combining reference types Mix videos and images in the same request to pull different qualities from each: motion and camera work from a clip, a subject's appearance from a still. Videos go in the same inputReferences array as images, tagged with a video media type. Numbering runs p

Marketplace integrations now install provider skills

August 6, 2026

When you install a Vercel Marketplace integration from the Vercel CLI, it now also installs that provider's agent skills from skills.sh, so your agents know how to use it: This happens automatically for any provider that publishes skills. You can also find integrations without leaving the terminal: vercel integration discover browses available providers. vercel integration categories lists all the Marketplace categories they're grouped into, like databases, AI, and observability. Update to the latest Vercel CLI with npm i -g vercel@latest, then read the integration documentation to get started. Read more

Agent Plugins 1.0.0: An Open Standard

August 6, 2026

Vercel announced Agent Plugins 1.0.0, an open, vendor-neutral standard for packaging AI agent extensions: a directory with a plugin.json manifest and fixed locations for Agent Skills and MCP servers. It's backed by representatives from AWS, Anysphere, GitHub, Microsoft, OpenAI, and Vercel, and already supported by ChatGPT, Codex, Cursor, GitHub Copilot, Kiro, and VS Code.

Measure time between steps in Vercel Workflows

August 5, 2026

You can now measure the time between any two steps in the trace viewer for Vercel Workflows and the Workflow SDK. Select a step, hold Option on macOS or Alt on Windows and Linux, then hover another step to see a measurement line between them: Between sequential steps, the line measures from the end of the earlier step to the start of the later step. Between overlapping or nested steps, it measures the difference between their start times. With no step selected, hold the modifier to see the gaps across the timeline, now drawn with the same measurement line. This can be useful for spotting delayed starts and queue waits when a run takes longer than its steps account for. To get started, open a workflow run from the Workflows tab in Vercel Observability, or locally with npx workflow@beta web. Read more

Full Sandbox egress firewall now available on Hobby plan

August 5, 2026

All Vercel Sandbox firewall features are now available on the Hobby plan. This brings the same network isolation that protects production workloads to the free tier, giving Hobby builders control over exactly what leaves the sandbox while keeping secrets out of the code entirely. Because the firewall attaches secrets to outbound requests itself, sandboxed code can call authenticated services like AI Gateway without ever seeing the token. Define allow-all, deny-all, or custom network policies with domain and IP-based rules, using matchers to scope each rule by path, method, query string, or headers. Custom policies can safely broker credentials or proxy requests through infrastructure you control. To get started, pass a networkPolicy when you create a sandbox. The Sandbox CLI manages the same policies, letting you update them live without restarting the sandbox. Network policies reduce data exfiltration risk when running untrusted or AI-generated code. Learn more in the Sandbox firewall documentation. Read more

AI Gateway is now available on AWS Marketplace

August 5, 2026

AI Gateway is available on AWS Marketplace. Teams can procure AI Gateway through their existing AWS account, consolidating inference spend onto their AWS bill and streamlining procurement. Purchases are available as private offers with annual contract terms, plus usage-based pricing beyond the contract. AI Gateway gives you one API to hundreds of models through a single endpoint. Reliability, cost controls, and governance, including automatic fallbacks, regional inference, and Zero Data Retention, are built in. You pay provider pricing with no markup; buying through AWS doesn't change what you pay per token. View the listing on AWS Marketplace for details. Read more

iMessage support for eve agents

August 5, 2026

eve agents can now connect to iMessage through the new Photon channel. The channel keeps each iMessage conversation in one eve session, verifies incoming webhooks, and marks accepted messages as read. If more messages arrive while the agent is replying, they steer the in-progress reply instead of each getting a separate answer. The onMessage hook filters which messages the agent handles and adds context to a turn. Run eve add channel/photon-imessage to add the channel. iMessage delivery runs through Photon, so the command sets up a Photon project, registers your phone number, and scaffolds the channel: eve uses Vercel Connect to create the connector and configure Photon's webhook. On hosts without Connect support, choose Use portable credentials during setup and the channel reads its credentials from environment variables instead. Read the Photon channel docs to get started, or deploy an agent from a template. Read more

Muse Spark 1.2 is now available on Vercel AI Gateway

August 5, 2026

Muse Spark 1.2 from Meta is now available on AI Gateway. It is a coding-focused update to the previous Muse Spark model. While keeping its general capabilities, 1.2 ships with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows. The model is built for long-horizon work like generating whole repositories, building out large projects end to end, and sustaining iterative loops where it writes, compiles, profiles, and improves code over many rounds. To use Muse Spark 1.2, set model to meta/muse-spark-1.2 in the AI SDK: To use Muse Spark in a coding agent, run vercel ai-gateway coding-agents setup to connect Claude Code, Codex, OpenCode, or Pi, then select meta/muse-spark-1.2 inside the agent. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. T

Vercel Sandbox now supports 10,000 concurrent sandboxes and 5,000 vCPUs per minute

August 5, 2026

We've increased the default quotas for Vercel Sandbox on Pro and Enterprise plans: Quota Before After Concurrent sandboxes 2,000 10,000 vCPUs allocation rate 200/min (Pro), 400/min (Enterprise) Up to 5,000/min The higher vCPU allocation rate is possible thanks to a new dynamic quota: instead of a fixed limit, the allowed rate now ramps with sustained usage: it starts at 150 vCPUs per minute and increases by 500 vCPUs per minute as you continue creating sandboxes, up to 5,000 vCPUs per minute The new quotas apply automatically to all Pro and Enterprise teams. If you need higher quotas, contact sales. Learn more about Sandbox pricing and dynamic quotas. Read more

Search and buy domains in the dashboard

August 5, 2026

You can now search for and purchase domains from the Vercel dashboard. Open the domain search panel from the Buy button in domain settings or the + button on the project overview. Nameservers are configured automatically for your project, so there are no DNS records to set. Read the working with domains documentation to get started. Read more

Set your own project avatars

August 5, 2026

You can now set a project's avatar from its settings in the Vercel dashboard. Until now, Vercel set it automatically, using the favicon from the project's production deployment. That's still the default if you haven't set one. The avatar renders across the dashboard and on Vercel bot comments in GitHub. Microfrontend projects that share one favicon can now get distinct avatars, and backend projects with no favicon to detect can set one explicitly. You can change a project's avatar from its settings in the Vercel Dashboard, or upload one with the REST API. Read more

New setup page after domain checkout

August 5, 2026

Buying a domain on Vercel now takes you to a setup page that tracks registration live, with direct paths to deploy a new project, connect an existing one, proxy or redirect a site you already run, or set up email. Registration can take a few minutes. Each setup action unlocks as soon as the domain is ready. If registration fails, the page shows what went wrong and the status of your refund. Deploy a new project or connect an existing one Deploy something opens project creation with the domain preselected. Connect a Git repository, prompt with v0, or start from a template, and the domain is attached when the project deploys. Connect an existing project attaches the domain to a project you pick from your team. Proxy or redirect a site you already run Enter an origin to proxy traffic to, and requests to the domain route through Vercel's CDN. Vercel will cache applicable requests at the edge, and Vercel Firewall's automated DDoS protections will automatically run. Enter a URL to redirect, and visitors are forwarded there. In both cases, Vercel creates a project to handle the routing and configures it for you. Set up email with DNS presets Choose your email provider and Vercel adds the

Export AI Gateway traces with Vercel Drains

August 5, 2026

AI Gateway now produces an OpenTelemetry trace for every request. Pro and Enterprise teams can send these traces through Vercel Drains to any OTLP/HTTP-compatible endpoint, including native integrations for Braintrust, Dash0, Kubiks, Sentry, and Statsig. Each trace shows the full request lifecycle, including: Model and provider routing Fallback and retry attempts Token usage and cost Time to first token, request duration, and response status Project, deployment, API key, environment, and custom tag attribution Trace Drains do not include prompt or completion content. Sampling controls let you choose how much traffic to export to each drain. AI Gateway traces cost $0.05 per 1,000 traces delivered to each drain, plus the standard Drains rate of $0.50 per GB of data transferred. A request counts once per successful drain delivery, even when AI Gateway makes multiple provider attempts. Failed deliveries do not incur a trace charge. Set up a Trace Drain from your team's Drains settings or learn more in the AI Gateway documentation. Read more

Introducing the new v0 API

August 5, 2026

Today we're introducing the new v0 API: programmatic, headless access to v0's app-building agent. Send a prompt and v0 generates an app, starts a dev server in a Vercel Sandbox, and gives you a preview URL you can embed in your own UI. Each chat is an isolated workspace for one app, where v0 can read, edit, and run the files. Follow-up messages continue from the current state. v0 verifies the code running in the Sandbox, so it can catch and fix errors in your app in real time. The new API is now generally available. How it works Create a chat from a prompt, stream its work, and render the live preview, all from your own interface. First, install the v0 SDK. Then, create a chat. Send a prompt, and v0 generates your app. Send follow-up messages to the same chat to keep building. Each one continues from the current state. Then embed the app. A prompt goes in, a running app comes out. v0 operates the Sandbox, dev server, and preview behind it. When you're ready, deploy to Vercel in one API call. v0 manages the whole pipeline. What this unlocks The new API lets you run v0's app builder inside your own product or pipeline. This opens use cases like: A white-labeled app builder: Your user

Vercel Sandbox now supports Devin Outposts

August 4, 2026

Vercel Sandbox now supports Devin Outposts. Each Devin session executes in its own isolated Sandbox microVM, with no local Outpost worker to keep online. The Devin control plane stays with Cognition, where the agent loop handles inference and planning. Session orchestration and command execution run in your Vercel project, and a durable workflow manages each session's lifecycle. With the integration, teams can: Preserve session state with Sandbox snapshots. Apply Devin network policies through Sandbox firewall rules. To get started, deploy the quickstart and approve the connection as a Devin administrator. Encrypted credentials are added to your project automatically, with no token or API key to copy. Then start a session in Devin and select the new Vercel outpost as its virtual environment. The integration is available to Pro and Enterprise teams with Devin Outposts enabled. Learn more in the knowledge base guide. Read more

Next.js 16.3 support on Vercel

August 4, 2026

Yesterday the Next.js team announced the release of Next.js 16.3, with leaner prefetching, immutable static assets, and instant navigations. As part of this release, we worked with the Next.js team to fully support 16.3 on Vercel, including better performance and additional observability. Applications that have upgraded have seen: Fewer prefetch requests: Apps upgrading from 16.2 to 16.3 saw 45% fewer prefetch requests on average, with some seeing reductions over 70%. Lower static asset costs: With immutable static assets, upgraded apps saw 17% fewer CDN requests and 24% fewer bytes transferred for static content, and up to 60% global TTFB reductions for frequently deployed projects. Faster routing at scale: Routing metadata improvements in the critical serving path deliver ~2x faster p99 route resolution for large sites. Platform improvements Next.js 16.3 ships leaner prefetching and enables immutable static assets by default. Supporting both meant changing the platform too, in how it caches static assets, resolves routes, and surfaces what happened in production. Immutable static assets Vercel now supports reusing immutable static assets across deployments for Next.js. Immutable

Deploys are now up to 33% faster for apps with many ISR pages

August 4, 2026

Deploys are now up to 33% faster for apps using Incremental Static Regeneration (ISR). The more prerendered pages you have, the more you're likely to save. ISR combines the speed of static generation with the flexibility of dynamic rendering, letting sites update content without full rebuilds. Every ISR page has routing metadata that ships alongside it. Previously, when that metadata grew too large, it split into a separate upload, adding work to every deploy. Now it always ships in a single upload. The same holds for other prerendered pages, including Partial Prerendering. This applies automatically to every deploy, with no configuration needed. Learn more about builds in the documentation. Read more

Give your eve agent a browser

August 4, 2026

Your eve agent can now navigate the web like a human with agent-browser. The @agent-browser/eve extension gives any eve agent a full set of browser tools: navigate pages, read content, click, fill forms, take screenshots, and inspect console and network activity. Everything runs inside the agent's sandbox. Install the extension: Then mount it under agent/extensions/: Your agent gets namespaced tools like browser__navigate, browser__snapshot, browser__click, browser__fill, and browser__screenshot. Snapshot refs such as [ref=e12] become @e12 selectors, so the agent can inspect a page, then act on exactly what it saw. The extension includes: Sandbox isolation: Chromium and agent-browser run in the eve sandbox, not your app. Pre-install them in your sandbox template so sessions start warm, or let the extension install them on first use. Domain allowlist: Restrict navigation and subresources with allowedDomains, and control output size, screenshot handling, proxy use, and session naming through extension options. Credential protection: Cookie, storage, and saved-auth-state commands are never exposed to the model. Use eve extension overrides to disable tools, require approvals, or add ap

DeepSeek V4 Flash is 90% off through Novita on AI Gateway

August 4, 2026

DeepSeek v4 Flash is 90% off on AI Gateway when you route to Novita. The discount is available to Vercel Pro customers through August 11. To get the discounted rate, set the model to deepseek/deepseek-v4-flash or deepseek/deepseek-v4-flash-0731 and put Novita first with the order option: If Novita can't serve the request, it falls back to other providers at the standard rate. After August 11, the model stays available at standard rates with no markup. Create an API key in the AI Gateway section of your dashboard, or try the model first in its playground. Read more

Vercel WAF for Blob is now generally available

August 3, 2026

Vercel WAF for Blob is now generally available and supported for production use on all plans. If you protected a store during the beta, nothing has changed: your rules and setup carry over exactly as they are. Vercel WAF protects a Blob store with custom rules, the same kind you use to guard your deployments. Rules that match on properties like IP address, country, and path can deny, challenge, or rate limit requests to your stored objects. This requires no changes to your code, your blob URLs, or how you use @vercel/blob. Rules are evaluated at Vercel's edge, so a denied request is turned away before any data transfer occurs. Unwanted traffic to AI-generated media, user uploads, or paid downloads never reaches your store and never generates a bill. To enable protection, open your Blob store's Settings tab, and in the Firewall section select Protect your store, then Enable firewall. This connects the store to a dedicated team-wide project where your Blob firewall rules live, shared by every store you protect. Read the Blob security documentation to protect a store. Read more

How Factory scaled its cloud backend to one billion monthly requests on Vercel

August 3, 2026

Factory on Vercel One billion backend API requests served daily 350ms p95 response time Scaled backend, internal tooling, and security without a dedicated infrastructure team Factory's mission is to bring autonomy to every phase of the software development lifecycle, from signals to production. Not to replace engineering judgment, but to automate the repetitive work around it, giving engineers more time to focus on the decisions that matter most. To deliver this enterprise-grade platform, Factory adopted their ideal operation model internally: remaining lean and agile by deploying Droids to manage routine workloads, so their engineers could focus strictly on building the product. As Factory scaled from a simple web app into a multi-surface platform, serving everyone from individual developers to enterprises with tens of thousands of engineers, their backend scaled alongside them. Today, a single Next.js backend on Vercel powers every surface, handling one billion daily requests across API routes, middleware, and webhook handlers, without becoming a separate engineering project. Next.js as a full-stack backend for cloud SaaS: API routes, middleware, and backend logic at scale Most t

Share Vercel Container Registry repositories across teams

August 3, 2026

Vercel Container Registry now lets you share a repository with other Vercel teams to grant read access to its images. Sharing is read-only and covers the whole repository. Recipients can pull and use every image, but cannot push, delete, or re-share. You can also share a repository with your own team. Each repository can be shared with up to 100 teams. Share from the project dashboard under Images → repository → Settings → Repository Sharing, or with the Vercel CLI: # Share a repository with a team vercel vcr permissions my-repository add other-team # List teams with access vercel vcr permissions my-repository ls # Revoke a team's access vercel vcr permissions my-repository rm other-team # Remove all shared access vercel vcr permissions my-repository clear Shared images can be used in Vercel Sandbox, accepting a team-scoped image reference in Sandbox.create() : import { Sandbox } from '@vercel/sandbox'; const sandbox = await Sandbox.create({ image: 'other-team/their-project/shared-repository:latest', }); Repository sharing is available on all plans. Update the Vercel CLI to v58.3.0 or later with npm i -g vercel@latest to get started and see the Vercel Container Registry documentati

Qwen 3.8 Max now available on Vercel AI Gateway

August 2, 2026

Qwen 3.8 Max is now available on AI Gateway. Qwen 3.8 Max handles text-only and vision-language work in one model, with 2.4 trillion parameters and a context window of up to 1 million tokens. The model is suited for software engineering and office productivity, along with visual work like turning screenshots or design files into working pages, captioning video, and answering questions grounded in an image. To use Qwen 3.8 Max, set model to alibaba/qwen3.8-max. Try Qwen 3.8 Max in the model playground. To use it in a coding agent, run vercel ai-gateway coding-agents setup to connect Claude Code, Codex, OpenCode, or Pi, then select alibaba/qwen3.8-max inside the agent. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Read more

Chat SDK now supports reactions and ephemeral messages on Teams

July 31, 2026

Chat SDK's Microsoft Teams adapter now supports reactions and ephemeral messages through the same API as other adapters. Bots can add and remove reactions on Teams messages, and call thread.postEphemeral() or channel.postEphemeral() to send native, targeted messages that only the intended user sees, like permission prompts. This release also adds: Author emails: Incoming Teams messages include the author's Microsoft Graph email, with their user principal name as a fallback. Custom token factory: Pass a token function in TeamsAdapterConfig to supply your own access tokens, so bots on runtimes without Azure IMDS can authenticate without a static client secret. Native DM status: Progress messages now show as a native Teams status in direct messages. Conversation routing: Group conversations are no longer misidentified as direct messages, so they're always handled correctly. Mention conversion: Email addresses, @handles inside URLs, and mentions inside code spans are no longer turned into Teams mentions. Read the Teams adapter documentation to get started, or learn more about ephemeral messages and reactions. Special thanks to onmax for their contributions to this release. Read more

AI Gateway logs now have a dedicated page

July 31, 2026

AI Gateway now has a dedicated Logs page listing every request your team sends through the gateway, newest first, with cost, token counts, duration, and the model, provider, and region that served it. Open any request to see how it was routed. It's available at the team scope and per project. Filter and search Filter by provider, model, modality, credentials, or status, or search by model, provider, or request ID to jump straight to one request. Every filter is stored in the URL, so you can share a filtered view by copying the address. A chart above the table shows request counts over the selected range, and you can drag across it to zoom into a narrower window. You can also export the current view as CSV or JSON, with your filters applied. Drill into a request Click any row to open its details next to the list: total cost, a token breakdown across input, output, reasoning, and cache reads and writes, duration, time to first token, and whether Zero Data Retention and region restrictions were applied. The fallback path lists every provider attempt for the request in order. Each attempt shows the model, provider, status code, credentials, and how long it took, and failed attempts inc

Vercel MCP now supports the 2026-07-28 MCP specification

July 31, 2026

Vercel MCP now supports the 2026-07-28 MCP specification, giving newer clients a stateless request model and updated authorization behavior without any change on the client side. Clients built for the 2025 protocol keep working exactly as before, and clients that understand the new specification pick it up automatically. Both protocol versions are served from the same endpoint through the official MCP SDK v2 and mcp-handler 2.x, which added support for the latest spec, so a single connection handles either protocol. Nothing about setup changes. Add the server the same way you always have. npx add-mcp https://mcp.vercel.com You can also run vercel mcp. See the Vercel MCP documentation to connect your client. For the full list of protocol changes, read the 2026 specification changelog and the authorization announcement. Read more

Expanded search for workflow runs in Vercel Observability

July 31, 2026

The runs list for Vercel Workflows in Observability now has structured search, giving you more ways to find and investigate runs. Structured filters: Filter runs by workflow, environment, deployment ID, region, or any custom run attribute. Paste a run ID to pull up that exact run. Smarter suggestions: The search bar suggests filters and values from your project's run data as you type, and recent searches are saved per project. Editable queries: Applied filters appear as removable chips that combine in one query, and the query lives in the URL, so a filtered view can be shared as a link and restores on load. For example, workflow:sendEmail environment:preview region:iad1 pulls up preview runs of one workflow in one region, since the list otherwise shows production runs. Search your runs from the Workflows tab in Vercel Observability. Read more

10x more capacity for Laguna S 2.1 on AI Gateway

July 31, 2026

Laguna S 2.1 from Poolside now has 10x more capacity on AI Gateway. The increase applies to the paid version, poolside/laguna-s-2.1, and the free version, poolside/laguna-s-2.1-free, so you can send far more requests, good for high-volume agentic coding and long-running tasks. To use Laguna S 2.1, set model to poolside/laguna-s-2.1-free or poolside/laguna-s-2.1 in the AI SDK: To run it in a coding agent, use vercel ai-gateway coding-agents setup to connect your agents to AI Gateway, then select poolside/laguna-s-2.1 or poolside/laguna-s-2.1-free in the agent's model configuration. See the coding agents guide. AI Gateway gives you one API to hundreds of models, with usage tracking, retries, failover, and higher-than-provider uptime built in. It reflects provider pricing with no markup and no platform fee, including on Bring Your Own Key requests. Try Laguna S 2.1 in the model playground. Read more

Vercel Passport is now generally available

July 31, 2026

Vercel Passport is now generally available. Passport allows you to protect your Vercel deployments with your own identity provider. Visitors authenticate through Okta, Microsoft Entra ID, or any OIDC provider before viewing a protected deployment, and Vercel forwards a signed identity token to the deployment so application code can build on who the visitor is. Read visitor identity in application code The getIdentity() helper in @vercel/passport reads the Vercel request context and returns the authenticated visitor. Vercel strips client-supplied values for the x-vercel-oidc-passport-token header and injects the verified token after Passport validates the session, so the identity your code receives is already verified. The subject field is a stable identifier for the visitor, scoped to your team and the Vercel Connect application that links Passport to your identity provider, and externalSubject is the visitor's ID in the provider itself. The helper returns null only when a request arrives without a Passport session, because unauthenticated browser visitors are redirected to the identity provider before they ever reach your code. In local development, getIdentity() returns a configu

AI Gateway adds unified fast mode support

July 31, 2026

AI Gateway has a unified fast mode abstraction, now in beta. You can now request fast mode the same way for every model on AI Gateway. Set speed to fast, and the gateway serves the fast tier when it's available and falls back to standard speed when it isn't. Fast mode trades a higher per-token cost for lower latency or higher throughput. You get it wherever it is available, without pinning a provider or wiring anything up yourself. Requesting fast mode Set speed to fast under providerOptions.gateway to upgrade the model to its fast serving path when one is routable: You can also address a fast variant directly with its fast slug, such as anthropic/claude-opus-5-fast, which is the same as setting speed: 'fast' on the base model. Use the speed option when you want to stay on the base model ID and fall back to standard speed. Use a fast slug when you want to name the fast variant directly, for example in a gateway.models fallback list. The speed option works across every AI Gateway API format. Fast mode is available for a growing set of models. Find them in the models list with a lightning bolt next to the slug. Requesting fast mode on a model with no fast tier has no effect, and the

AI Gateway now supports team and project spend budgets

July 31, 2026

AI Gateway spend budgets can now be scoped to a team or a project, in addition to individual API keys. Set a dollar limit on a scope, and the gateway meters spend against it and stops further requests once the limit is reached, until the budget resets or you raise it. Scopes A budget attaches to one of three scopes. A request can fall under multiple budgets and has to pass all of them: if any one is over its limit, the request is rejected, even when the others have room left. Scope Caps spend for Team Every request your team runs through the gateway Project All requests attributed to a single project API key A single API key BYOK spend is not counted against budgets by default. Manage budgets in the dashboard Set and track budgets from the AI Gateway Budgets tab. The Overview breaks spend down across every scope against its limit, so you can see what is near or over before it starts rejecting requests. Turn on email spend alerts to notify your team's usage notification recipients when usage crosses 50%, 75%, or 100% of a limit within a refresh period. Alerts are off by default and informational: they never block requests, only the budget limit does. Set a default budget for project

DeepSeek V4 Flash now runs updated weights on AI Gateway

July 31, 2026

DeepSeek V4 Flash now runs on updated weights by default on AI Gateway, with notably stronger agentic capabilities. On Terminal-Bench, it scores 82.7, up 25.8 points from 56.9 in the April preview. Requests to deepseek/deepseek-v4-flash pick up the new weights automatically, with no change to the model ID or your code. For now, DeepSeek is the only provider serving the updated weights. Other providers, including ones with Zero Data Retention, are coming next week. To use the updated DeepSeek V4 Flash, set model to deepseek/deepseek-v4-flash in the AI SDK. AI Gateway will route to providers with the new weights by default: To run V4 Flash in a coding agent, use vercel ai-gateway coding-agents setup to connect your agents to AI Gateway, then select deepseek/deepseek-v4-flash in the agent's model configuration. See the coding agents guide. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Run background tasks with Celery on Vercel

July 31, 2026

Celery, an asynchronous distributed task queue for Python, is now supported natively on Vercel. Tasks are executed as Vercel Functions and automatically scale with traffic. By default, task results are stored in Runtime Cache, which is sufficient for small data sizes and relatively short workflow runtimes. For workloads that require stronger persistence guarantees, more storage or longer retention, a durable result backend can be configured. Celery workers can be declared as subscribers in pyproject.toml. When running on Vercel, the vercel:// broker is automatically installed and to use Vercel Queues, and the default Celery broker_url is set to vercel://. Celery workloads on Vercel use Fluid compute with Active CPU pricing by default. This means your Celery workers will automatically scale up and down based on traffic, and you only pay for what you use. Deploy Celery workloads on Vercel or visit the Python runtime documentation. Read more

Deployments are now up to 7 seconds faster

July 30, 2026

Over the past few months, we've made deployments up to 7 seconds faster end to end. We removed up to 5 seconds of fixed platform overhead from every build, and deploys from the latest Vercel CLI save up to 2 seconds more. The improvements are most noticeable on smaller builds, where orchestration represents a larger share of the total build time. The largest platform gains: ~2.2s: Internal build-process shutdown now happens outside the critical path to deployment readiness ~930ms: The Vercel CLI is prepared ahead of time inside the secure, isolated build environment, removing startup work from the critical path ~413ms: Deployment finalization starts earlier, processing routing and output metadata concurrently with other work ~900ms: Redundant API and storage requests are eliminated across build start and finalization, including fetching deployment configuration concurrently with the build and loading build records directly by ID These apply automatically to builds triggered through Git, the dashboard, or Vercel CLI. Deployments from the latest CLI finish up to another 2.1 seconds sooner: ~1s: The CLI completes as soon as the event stream reports domain aliases are assigned, instead

Latest MCP spec now supported in mcp-handler

July 30, 2026

mcp-handler@2.0.0 is now available with support for the 2026-07-28 Model Context Protocol specification and MCP TypeScript SDK v2. We originally built mcp-handler to make it easier to spin up MCP servers on popular web frameworks including Next, Nuxt, Svelte, and more. With the new 2.0 release, the handler now supports: The stateless 2026-07-28 protocol, served natively, including per-request metadata and server/discover A stateless compatibility layer for clients using 2025-era Streamable HTTP Both protocols with no Redis dependency or session storage Existing stateless Streamable HTTP clients can continue connecting to the same /mcp endpoint while servers adopt the latest protocol. Install the new packages: Create a Next.js route handler: MCP clients can connect to /api/mcp. Get started with the Next.js MCP server template, updated for mcp-handler 2.0. Before you upgrade Version 2 requires Node.js 20 or later and zod4, and replaces @modelcontextprotocol/sdk with @modelcontextprotocol/server. Tool, prompt, and resource definitions now use the SDK v2 registration APIs and Standard Schema inputs, such as registerTool with z.object(...). The deprecated HTTP+SSE transport has been rem

Introducing Enterprise Flexible Commitment for Vercel Marketplace

July 30, 2026

Enterprise customers can now apply a portion of their Flexible Commitment toward eligible resources purchased through the Vercel Marketplace. With Flex Commit support, eligible Marketplace cost can draw directly from your existing commitment making it easier to provision the infrastructure and services your applications need. We're rolling this out with an initial set of Marketplace partners, including Neon, Supabase, Mongodb and Redis, with support for more providers coming. Eligible integrations are marked with a Flexible Commitment Eligible badge on their Marketplace page. With Marketplace Flex Commit you can: Use committed Enterprise spend for eligible Marketplace resources Eliminate separate credit card or procurement workflows Provision and manage third-party services directly from the Vercel Marketplace Track Marketplace usage alongside the rest of your Enterprise commitment This is the first step toward making the Marketplace a seamless part of the Enterprise experience. To learn more, read our pricing page documentation and reach out to your Account Manager. Read more

Project-scoped Tokens

July 30, 2026

You can now create Vercel Access Tokens that are limited to a project to authenticate and use the Vercel API. A project-scoped token can only read and write resources belonging to a project that the token is scoped to. Requests to any other project, a user-level resource, or a team-level resource will be denied. This ensures jobs, tools, or workflows only ever access the projects they are scoped to. Creating a project-scoped token Navigate to the Account Tokens page, found under the Settings area of your Account. Enter a descriptive token name Open the Scope dropdown and select the team that owns the project, then click the team to drill into its list of projects Select the project you want the token to be limited to Choose an expiration and click Create Make a note of the token created as it will not be shown again Read more

Server-Timing response headers will pass through to the client

July 30, 2026

On August 10, 2026, Vercel's CDN will stop stripping the Server-Timing response header and begin passing it through to the client. Use Server-Timing to report backend metrics like database query time and cache hits. These values appear in the browser's network panel and as PerformanceServerTiming entries through the Performance API. If you want to keep the current behavior, where the header is stripped from every response, you can add a transform in vercel.json: Learn more about the Performance API and Server Timing. Read more

Shopify and Vercel are rebuilding Hydrogen for faster storefronts

July 30, 2026

Shopify and Vercel Open source and runtime agnostic, runs anywhere JavaScript does Skills let agents extend any storefront with a single command Feature development cut from months to a week for retailers like Paige Shopify powers commerce for millions of merchants worldwide, and Hydrogen is its framework for building headless storefronts. The Shopify team is partnering with Vercel to rebuild Hydrogen from the ground up. The new version is open source and runtime agnostic so developers can build with Next.js, Nuxt, Svelte, or any framework they choose. As a result, storefronts that once required custom glue code now ship with best practices built in, from type-safe API clients to cart state in a single import. Read more

Run multiple isolated agents in a single Sandbox

July 30, 2026

The @vercel/sandbox SDK now supports multiple Linux users and groups, so you can run agents side by side in a single Sandbox. Each agent runs as its own user with a private home directory. A group opens a shared workspace when they need to collaborate. This makes multi-agent systems easier to build. Call createUser for each agent; its commands and file operations run as that user, and users can't read, write, or list each other's files. To set up a shared directory, call createGroup and add users to it. Learn more in the Vercel Sandbox documentation. Read more

MiniMax H3 now available on AI Gateway

July 30, 2026

MiniMax H3 is now available on AI Gateway. H3 generates 2K video from a text prompt, a starting image, a pair of first and last frames, or reference material. Alongside text-to-video and first-frame image-to-video, the model supports first-to-last keyframe transitions and multimodal reference-to-video, conditioning a generation on reference images, video, or audio in a single request. Reference and keyframe modes are mutually exclusive. Output is mp4 at 2K resolution, from 5 to 15 seconds, in aspect ratios including 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16, or adaptive to a supplied image. To use MiniMax H3, set model to minimax/minimax-h3 in the AI SDK: Pass a starting image with the prompt, and the model animates it following the image's aspect ratio. Supply reference images, video, or audio as source material, and refer to each by its order in the prompt for reference to video output. View all video models on AI Gateway. Try MiniMax H3 in the model playground. Read more

Inkling Small from Thinking Machines is now available on AI Gateway

July 30, 2026

Inkling Small from Thinking Machines is now available on AI Gateway. Inkling Small reaches performance comparable to the larger Inkling model at about a quarter of the size, using much less compute per task. It is a broad generalist with native reasoning over audio and images, and it holds up well on reasoning, agentic coding, and tool use. Controllable thinking effort lets you trade quality against cost and latency, from minimal to maximum reasoning. For visual tasks, it can crop, zoom, and inspect images programmatically, which helps on documents and charts where the relevant detail is small. To use Inkling, set model to thinkingmachines/inkling-small in the AI SDK: Inkling-Small is compatible with Zero Data Retention. Turn it on team-wide from the dashboard, or per request with zeroDataRetention: true, and AI Gateway routes only to providers that delete prompts and responses after each request. Inkling-Small is also a cost-efficient choice for coding and tool-use workflows. Run vercel ai-gateway coding-agents setup to connect your coding agents to AI Gateway, then select thinkingmachines/inkling-small in the agent's model configuration. See the coding agents guide. AI Gateway re

Turborepo and Vercel Remote Cache now support OpenID Connect (OIDC)

July 30, 2026

You can now exchange OIDC tokens from CI/CD workflows, including GitHub workflows, for short-lived Turborepo access tokens. These tokens grant access to Vercel's Remote Cache, and are a more secure alternative to long-lived Personal Access Tokens (PATs). OIDC tokens are short-lived, only grant access to Vercel Remote Cache, and are associated with your Vercel team, rather than a specific team member. We recommend all customers migrate their CI/CD workflows from PATs to OIDC. Get started by adding a Turborepo OIDC policy under Settings → Build and Deployment → OIDC Policies for CLI Access and learn more in the documentation. Read more

AI Gateway: GPT

July 30, 2026

On AI Gateway, GPT-5.6 Luna and GPT-5.6 Terra are now cheaper and GPT-5.6 Sol is faster. AI Gateway adds no markup on token pricing, so these changes reach you at the upstream rate. The changes apply to both short and long context pricing. Model Change Input: Short context (per 1M tokens) Output: Short context (per 1M tokens) openai/gpt-5.6-luna 80% price reduction $0.2 $1.2 openai/gpt-5.6-terra 20% price reduction $2 $12 openai/gpt-5.6-sol Same price, fast mode now 2.5x faster (up from 1.5x) Unchanged Unchanged GPT-5.6 Sol keeps the same price; its fast mode now runs 2.5x faster, up from 1.5x. Model IDs are unchanged, so existing requests get the new rates and speed with no code change. See full pricing detail on AI Gateway. Read more

Additional custom environments can now be purchased

July 29, 2026

Pro and Enterprise teams can now purchase additional custom environment capacity without contacting sales. Custom environments let you model your team's release process on Vercel. You can add staging, qa, or any named stage between preview and production, and each environment gets its own branch tracking, environment variables, and domains. You can purchase or adjust capacity from the dashboard, API, or CLI. Dashboard Custom Environments API POST /v1/projects/custom-environments/settings CLI vercel buy addon custom-environment Capacity is sold in packs of 5 environments. You can also reduce capacity as long as the project isn’t using more environments than the new limit allows. Billing updates when you confirm a change. Pricing and limits Custom environments are available on Pro and Enterprise plans. Additional capacity can be purchased for $50/month per pack of 5 environments. Plan Limits Pro 1 custom environment included per project (up to 16 total per project) Enterprise 12 custom environments included per project (up to 22 total per project) Purchases require an active Pro or Enterprise plan, billing permissions, and a valid payment method. Capacity is configured per project, a

Discover and install eve integrations from the CLI

July 29, 2026

You can now discover and install integrations for eve agents directly from the eve CLI. Integrations come from the official eve catalog and third-party sources. Run eve add from your eve project to install an integration: Integrations write their files directly into your project and can add anything an eve agent uses, from a single tool to a channel to a full extension. Review the generated files and add any required configuration before running your agent. Find integrations with the new eve registry commands: eve registry list: List available integrations. eve registry search : Search the catalog for a capability, like browser. eve registry view : Inspect an integration before you install it. You can also browse the integrations directory to see the official catalog. Add third-party sources with a namespace and URL template: Then install from that source with eve add @acme/analytics. Registries use the shadcn registry format, so any compatible registry works. You can also pass an integration URL directly without configuring a source: Integrations can add dependencies and write files, so treat them like project code. Add sources you trust, inspect integrations with eve registry vie

Grok Voice Think Fast 2.0 now available on AI Gateway

July 29, 2026

Grok Voice Think Fast 2.0 from xAI is now available on AI Gateway. It is a speech-to-speech voice model that takes audio in and audio out, improving on the previous Grok Voice model in reasoning, transcription accuracy, and conversation. The model reasons in parallel with speech, so it can think through a query while talking without adding latency. It has also been trained to use fewer reasoning tokens than before, so tool calls fire sooner, often before the end of the agent's first sentence. Transcription holds up in real-world conditions, including background noise and telephony compression. Use xai/grok-voice-think-fast-2.0 through the AI SDK's realtime API. Mint a short-lived token on the server so your API key never reaches the client: See the realtime documentation to build a voice agent with Grok Voice Think Fast 2.0. Try out the model in the AI Gateway playground. Read more

Edge Config is now Global Config

July 29, 2026

Edge Config is now Global Config. This rename better reflects that it is a globally replicated data store with ~1ms reads in every region, built for the configuration that applications read at runtime, such as feature flags, redirects, and experimentation settings. Global Config stores can now hold up to 1 MB on every plan, and Pro and Enterprise teams can create unlimited stores with more writes per day. The rename also ships with a new server-side SDK, @vercel/global-config, and a new environment variable. What changed Name: Global Config is the new name for Edge Config. The store itself is unchanged. Package: @vercel/global-config replaces @vercel/edge-config as a drop-in replacement. Environment variable: Connecting a Global Config store to a project creates a GLOBAL_CONFIG environment variable instead of EDGE_CONFIG. Limits: Raised on every plan. Pricing is unchanged. What you need to do Existing projects: no action required. Deployments, connected stores, and the EDGE_CONFIG environment variable continue to work. Upgrading proactively is safe. The new SDK reads GLOBAL_CONFIG by default and falls back to EDGE_CONFIG, so it works with stores connected before and after the renam

Sign in with ChatGPT is now available on Vercel

July 29, 2026

Sign in with ChatGPT adds your ChatGPT account as an authentication option for Vercel. It is available when adding the Vercel plugin to ChatGPT and when signing in to Vercel or v0. When you add the Vercel plugin to ChatGPT, you can easily sign in to Vercel and grant team and project permissions without leaving ChatGPT. "Continue with ChatGPT" also appears as a sign-in option on the Vercel and v0 sign-in and sign-up pages. Team requirements like 2FA and SSO still apply. You can re-authenticate or disconnect the ChatGPT sign-in method in your account Authentication settings. Sign in with ChatGPT is in beta. Add the Vercel plugin to ChatGPT to try it and read OpenAI's changelog to learn more. Read more

Vercel Sandbox supports forking

July 28, 2026

Vercel Sandbox now supports forking with Sandbox.fork(). The fork starts from the source's current snapshot and inherits its config and environment variables. If the source is running, it forks the latest saved state, not the live in-memory state. If the source has no snapshot, it falls back to a fresh create, using the source's runtime and config. Any parameter you pass overrides the inherited value. A fork takes about the same time as creating a sandbox, with the same limits. Use it to branch an agent from a shared base, give each tenant its own copy of a template, or run variations of one setup in parallel. To fork a sandbox with the SDK: To fork a sandbox with the CLI: To get started, upgrade to the latest version: pnpm install @vercel/sandbox@latest # SDK pnpm install -g sandbox@latest # CLI Learn more in the Sandbox documentation. Read more

Vercel Connect now supports Custom Environments

July 28, 2026

Vercel Connect now lets you link a connector to a Custom Environment so deployments there can request provider tokens and receive forwarded webhooks. Custom Environments appear alongside Production, Preview, and Development when you add or edit a project on a connector in the dashboard. From the CLI, pass the environment's slug to --environment: Passing --environment replaces the default environments. Once linked, a deployment in that environment can request tokens with getToken as usual. If the calling environment isn't on the project link, the request fails with ClientNotEnabledForEnvironmentError. Trigger destinations can also target a Custom Environment: Before targeting a Custom Environment with triggers, deploy to it and assign it a verified domain. Vercel Connect forwards events through that domain. Project links control which deployments can request tokens, and they do not restrict a provider token after it is issued. When environments need provider-level isolation, create a separate connector for each environment and request only the scopes it needs. Vercel Connect is in beta on all plans. Custom Environments are available on Pro and Enterprise plans. Get started with the

DeepsecBench: evaluating model performance in finding cybersecurity vulnerabilities

July 27, 2026

Last week, OpenAI evaluated two models on an exploit benchmark within an isolated sandbox. Guardrails were reduced for testing, and the models found a vulnerability in their environment, accessed the internet, and reached Hugging Face's production database. No human directed the action, but the breach is a clear example of how much more capable malicious attackers are when equipped with powerful AI models. But defenders have the same tools, and a clear advantage: knowledge of their own codebase. Hacks are initiated from the outside, so the single best defense is finding vulnerabilities from the inside before attackers do. Today we're releasing DeepsecBench, a benchmark that evaluates how well different models find cybersecurity vulnerabilities in application code. For each model the report includes recall, precision, cost, and total time, and combines recall and precision into a single benchmark score. Here is a sample of model performance from the leaderboard: Rank Model Level Score Cost Total time 1 GPT-5.6 Sol xhigh 35.58 $55.98 03:39:00 3 Claude Opus 5 medium 28.36 $31.96 00:47:01 8 Kimi K3 high 17.56 $12.38 01:59:00 10 Grok 4.5 high 15.58 $5.60 01:24:00 We built deepsec to mak

WebSocket support for OpenAI Responses API live on AI Gateway

July 27, 2026

AI Gateway now supports WebSocket mode for the OpenAI Responses API. You can keep a persistent connection open and continue each turn by sending only new input items plus previous_response_id, instead of re-sending the full context over a fresh HTTP request every turn. OpenAI reports up to ~40% faster end-to-end execution on WebSockets for agentic rollouts with 20 or more tool calls. Route Responses API traffic over WebSocket The Responses route opens at GET /v1/responses and accepts raw response.create frames over a persistent WebSocket connection. The mode is compatible with store=false and Zero Data Retention, and follows OpenAI's WebSocket mode specification upstream. See the AI Gateway documentation for setup and the full route list. Read more

Nuxt July 2026 security advisory

July 27, 2026

The Nuxt team has released Nuxt 4.5.1 and 3.21.10, along with @nuxt/devtools 3.3.1, to address eight security advisories, including a high-severity server-side remote code execution vulnerability. Vulnerabilities addressed Vulnerability Severity Advisory Server-side remote code execution via server island props High GHSA-9473-5f9j-94wq Unauthorized component instantiation via server island props Medium GHSA-48hr-524c-v5w3 Route rule authorization bypass High GHSA-hxvh-4h3w-prp9 Server component denial of service High GHSA-hxcr-hm88-mpq6, GHSA-9pgf-384g-p7mv Cross-user disclosure of cached payloads, affecting Nuxt 4.x versions 4.4.0 and later High GHSA-wm8w-6qjm-cv43 Development server path disclosure Low GHSA-7c4v-fwgw-9rf7 Remote code execution in Nuxt DevTools, development-only, fixed in @nuxt/devtools 3.3.1 Critical GHSA-279x-mwfv-vcqv Vercel platform protections Vercel received advance notice of the server-side remote code execution vulnerability, GHSA-9473-5f9j-94wq, and deployed platform-wide WAF mitigations before public disclosure. Applications deployed on Vercel are automatically protected by these mitigations, with no configuration changes required. However, do not rely o

Kimi K3 and Kimi K3 Fast with ZDR and US-based providers now on AI Gateway

July 27, 2026

Kimi K3 from Moonshot AI and its faster serving path, Kimi K3 Fast, are now available from US-based providers on AI Gateway, including Baseten and Fireworks. Zero Data Retention (ZDR) is also supported for both models. Running Kimi K3 on US-based providers lets teams with data residency and compliance requirements use the model on US infrastructure. Because AI Gateway serves the models from multiple providers, it automatically routes across them for failover, higher uptime, and more available throughput than any single provider offers. You call the same moonshotai/kimi-k3 model ID, and the gateway handles provider selection and fallback. Kimi K3 Fast trades a higher per-token cost for lower latency. Request it with the speed option on the base model, which stays on moonshotai/kimi-k3 and falls back to standard speed when the fast tier is unavailable. Alternatively, use moonshotai/kimi-k3-fast. The fast variant costs ~50% more than the base model. To use Kimi K3, set model to moonshotai/kimi-k3 in the AI SDK: US inference To route Kimi K3 requests to use only US data centers for inference, set inferenceRegion. Regional pricing is ~10% more than the regular variant. Zero Data Retenti

Regional inference now available on AI Gateway

July 27, 2026

AI Gateway now supports regional inference. Set inferenceRegion on a request to pin it to the US or EU. Every model provider that supports the selected region handles it the same way. Inference runs there, and any data the provider keeps is stored there. AI Gateway supports two pinned regions, plus global routing: If no model provider can serve it, the request fails rather than running somewhere else. Every response reports the region that served it, so you can confirm where each request ran. Here's a request pinned to the US with the AI SDK: Until now, teams with data residency or compliance requirements had to configure regional routing separately for every provider, with no reliable way to confirm where a request actually ran. Regional inference replaces that with a single field that behaves the same everywhere and a response that tells you where each request was served. Filter the model list for models available in the US or EU, or read the regions array from /v1/models. Without inferenceRegion, requests route globally with no residency guarantee, so residency is opt-in. Pinning a region can cost more. The provider sets the regional rate, often around 10% above standard, and AI

eve adds new Slack event hooks and session controls

July 27, 2026

eve agents on Slack can now keep replying in a thread without repeated mentions, cancel an in-progress response or reset a conversation entirely, and react to any event your Slack app subscribes to. Continue conversations without repeated mentions Mentions no longer have to carry the conversation. Once a thread has an active session, your agent can reply on its own. The new onMessage hook receives incoming Slack messages, and two helpers decide which ones to handle: ctx.isBotMentioned() detects an explicit mention, and ctx.isSubscribed() checks whether the message belongs to a thread with an active eve session: When routing depends on who has joined the thread, the new ctx.thread.listParticipants() helper returns unique human Slack user ids in first-appearance order. Public channel messages require the message.channels trigger event type and the channels:history bot scope; private channels also need message.groups and groups:history. Cancel a turn or reset the conversation Message hooks also receive two thread-bound session helpers. ctx.cancel() stops the current turn while keeping the session: call it before returning { auth } to queue the new message as replacement input, so a mi

How Sandstone grew 40x in 147 days on Vercel

July 27, 2026

Sandstone on Vercel 1,000+ legal requests managed daily across customer teams 7-app turborepo monorepo deployed seamlessly to multiple Vercel projects Multi-step agentic legal workflows built end-to-end with AI SDK When in-house legal teams get a request from their business, it kicks off a manual process of pulling data from multiple systems. In a typical workflow, lawyers will pull deal details from Salesforce, review past contracts, and chase account teams for missing context, all before they start the legal work itself. Sandstone automates this data gathering process, and centralizes requests into a single platform. When a request is created, Sandstone automatically ingests context from every system, including messaging and email tools. By the time the lawyer sees it, they have all of the information they need to get to work. The founding team includes a McKinsey legal tech veteran, a lawyer-turned-engineer, and cybersecurity founder. At Sandstone, that mix shows up directly in the product: legal professionals work alongside engineers, catching the workflow details and legal nuances that only someone who has practiced law would notice. That is how Sandstone ships fast and correc

Vercel WAF for Blob is now in beta

July 24, 2026

The Vercel WAF can now protect a Vercel Blob store. The same rules that guard your deployments (deny, challenge, rate limit) now apply to blob traffic with no changes to your code, blob URLs, or @vercel/blob. Every blob is already served through Vercel's CDN, so protection is a switch on the store, not a new proxy. Stop scrapers, geo-restrict downloads, rate limit expensive assets, or block abusive IPs before a byte is served. Rules evaluate at the edge, matching on IP, country, path, and more: Deny returns a 403 and stops the request early, so no data transfer is incurred. Challenge serves the standard browser challenge, and a request that fails it is blocked. Rate limit returns a 429 when a client exceeds your limit. Redirect and log behave as they do for deployment traffic. The OWASP Core Ruleset is not supported, since it targets dynamic application traffic, not object delivery. Protecting a store Setup is a single switch in the dashboard: Open your Blob store, select Settings, then Protect your store. Vercel connects it to a shared vercel-blob-default-project on your team. You author its rules with the standard rule builder, and they take effect immediately. One rule set cover

Workflow steps now support extended function durations

July 24, 2026

Workflow steps on Pro and Enterprise plans can now run for up to 30 minutes (1800 seconds), up from 800 seconds, using extended function durations (in beta). To opt in, set VERCEL_ENABLE_WORKFLOW_EXTENDED_MAX_DURATION to 1 in your project's Environment Variables, then redeploy. Requires Fluid compute and a supported Node.js or Python runtime. Extended durations are not available on Hobby plans, which remain limited to 5 minutes (300 seconds). See function duration limits by plan in the documentation. Read more

Claude Opus 5 now available on AI Gateway

July 24, 2026

Claude Opus 5 from Anthropic is now available on AI Gateway. Opus 5 improves on previous Opus models for long-horizon agentic coding, handling multi-file features, larger refactors, and end-to-end feature work, and completing full tasks rather than leaving stubs or placeholders. Opus 5 is effective at low and medium effort, which produce quality at a fraction of the tokens and latency of higher settings. Vision is stronger on charts, documents, diagrams, and UI replication, and Opus 5 makes effective use of tools to analyze, crop, and verify visual work. It also coordinates teams of subagents well, which suits multi-agent workflows. Reasoning is on by default. In AI SDK 7, set the reasoning effort with the top-level reasoning option, from minimal up to xhigh, or none: Like Fable 5, Opus 5 ships with elevated cybersecurity safeguards, so benign security work can occasionally trigger a safety classifier and get blocked. To add a backstop, use AI Gateway's model fallbacks. List backup models in a models array under providerOptions.gateway, and the gateway tries them in order when the primary model fails. The same option works across every AI Gateway API format. Opus 5 is compatible wi

Ling 3.0 Flash is now available on AI Gateway

July 23, 2026

Ling 3.0 Flash from Ant Group is now available on AI Gateway. The model is free to use for the next three weeks, through August 3rd. Ling 3.0 Flash is a Mixture-of-Experts model with 124B total parameters and about 5.1B active per token. It has a 256K token context window and runs in thinking and non-thinking modes. Ling 3.0 Flash is built for token-efficient agentic inference at production scale, doing more work within tighter token, latency, and cost budgets across multi-step agent runs. The model targets high-frequency agentic workflows, coding agents, document work, and long-context multi-turn interactions. To use Ling 3.0 Flash, set model to inclusionai/ling-3.0-flash-free in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Ling 3.0 Flash in the model playground. Read more

Vercel MCP can now deploy code

July 23, 2026

The Vercel MCP server can now deploy code directly to a new or existing project. When your AI assistant finishes building something, it can ship it to Vercel and hand back a shareable URL without leaving the chat. Point the deploy_to_vercel tool at your files and Vercel creates the project, detects the framework, installs dependencies, and builds. You get a URL you can open and share while the build finishes in the background. To get started, connect the Vercel MCP server to Claude, Cursor, or any supported MCP client, and ask it to deploy your project. Learn more about the Vercel MCP. Read more

WebSocket support is now available for Python Functions

July 23, 2026

Vercel now supports WebSocket connections for Python applications. WebSockets enable bidirectional communication between client- and server-side code, powering real-time features like interactive AI streaming, real-time chat, and multiplayer live collaboration. Both ASGI and WSGI applications are supported, including frameworks like FastAPI, Django, and Flask. Explore the FastAPI AI Chat and Flask AI Chat examples, or read the WebSockets documentation to get started. Read more

Evaluation metrics for Vercel Flags

July 23, 2026

Vercel Flags now shows a live evaluation view on each flag's detail page. You can see evaluations per minute charted over time, with each flag version marked in the chart so you can tie evaluation shifts to specific configuration changes. You can group and filter evaluations by variant, reason, environment, SDK key, client, and reporting project. Custom clients can set the new clientName property on the client initializer to show up as their own group. This makes it straightforward to confirm rollout percentages match what's being served, separate targeted matches from default fallbacks, and see which projects are using a given flag. Upgrade to @vercel/flags-core 1.6.0 or later to get started. Metrics appear automatically with no additional configuration. Learn more in the evaluation metrics documentation. Read more

Connect to and manage Sandboxes from the dashboard

July 23, 2026

You can now connect to any running Sandbox from the Vercel dashboard. Run commands, browse the filesystem, upload and download files, and inspect open ports without leaving the browser. From the same view, you can also manage the sandbox lifecycle: take snapshots, and stop or resume persistent sandboxes. Open the Connect tab on any sandbox to try it out. Learn more about Vercel Sandboxes in the documentation. Read more

Inspect feature flag history with Vercel CLI

July 23, 2026

Vercel Flags version history can now be inspected from the Vercel CLI with the new vercel flags versions command. Run vercel flags versions to print the full revision history for a flag, with each revision's author, message, timestamp, and changed environments. Filter to a specific environment with --environment, paginate with --limit and --cursor, or add --json for scripting. To compare a revision against the one before it, run vercel flags versions diff --revision . The diff is semantic, surfacing field-level changes to targeting rules, rollout percentages, and conditions. Additions, removals, and modifications appear as colored +, -, and ~. Update to the latest version of the Vercel CLI and learn more in the Vercel Flags CLI documentation to get started. Read more

GitHub tools are now an installable eve extension

July 23, 2026

You can now add GitHub tools to your eve agent as an extension. Add the package, drop one file in agent/extensions/, and your agent gets every tool with Vercel Connect auth, presets, and approval rules built in. Install @github-tools/eve-extension: Then register it from a file in agent/extensions/: Connector-backed auth: Pass a Vercel Connect connector and the extension mints short-lived, scoped GitHub tokens at runtime. Presets scope the toolset: code-review, issue-triage, repo-explorer, ci-ops, and maintainer map to Connect scopes automatically, so tokens carry only the permissions the tools need. Approval rules travel with the config: Every write tool requires approval by default. Gate individual tools with always, once, or an input-dependent predicate, like approving comments only outside your own org. Namespaced and versioned: The filename sets the namespace, so tools run in the agent as github__addPullRequestComment. Upgrade the package to pick up new tools and fixes, and the config schema is validated on import. Get started by creating a GitHub connector and reading the documentation. Read more

AI Gateway now supports streaming transcription

July 22, 2026

AI Gateway now supports streaming transcription. Previously, transcription required a complete audio file and returned the full transcript in a single response. Now you can stream audio in as it's captured and get transcript updates back as the model produces them, keeping latency low for uses like live captioning and voice input. Streaming transcription is in beta and available through the AI SDK's streamTranscribe function with any streaming-capable transcription model. The example below streams raw PCM audio to openai/gpt-realtime-whisper and prints each transcript delta as it arrives. The result stream also carries partial and final transcripts: The same code works across providers: swap the model string to use xai/grok-stt or any other streaming-capable transcription model. Streaming transcription also makes it easy to add a voice mode to an agent. Stream the user's speech to a transcription model and pass the live text to your agent. The agent itself does not change: it still receives text, so this works with any text-based agent. For agents that speak back, pair it with speech generation, or use realtime voice for full two-way conversation. For more information on streaming

Extend eve agents with installable extensions

July 22, 2026

You can now package tools, connections, skills, instructions, and hooks into extensions that any eve agent can import. Extensions can be published to package registries, then installed, versioned, and upgraded like any other project dependency. A browser-use extension might ship tools for navigating a site, a memory extension can capture context with hooks and recall it with tools, and a self-improvement extension pairs hooks with dynamic instructions. Scaffold a new extension with a single command: This creates the package, installs dependencies, and initializes Git. The generated package is shaped like an agent, with tools, connections, skills, and hooks following the same file conventions. When the extension is ready, running eve extension build generates the publishable package. To use an extension, install the package and import it from a file in agent/extensions/: The filename sets the namespace, so the extension's search tool runs in the agent as crm__search. Extensions allow you to: Declare a config schema using a standard schema library, such as Zod. Consumer settings are validated on import and typed everywhere the extension reads them. Require approval before an extensio

Vercel Connect adds 100+ preset connectors

July 21, 2026

Vercel Connect now includes preset connectors for 100+ services, including Shopify, Resend, Sanity, Okta, Workday, and Jira. Preset connectors are predefined configurations for supported services. They reduce manual setup by pre-populating the brand name, icon, auth type, and MCP or discovery URL. Unlike managed connectors, preset connectors don't register your app with the external service for you. Add a connector Select a preset from the connectors directory in the Vercel dashboard. Review the pre-populated configuration, add any credentials the service requires, and create the connector. You can also start from the CLI: Today, vercel connect create opens the dashboard to complete setup. Full CLI support for preset connectors is coming soon. Request tokens After attaching a preset connector, request tokens the same way you do for any other Connect provider. Call getToken with the connector's UID (service/name) to receive a short-lived, scoped token: To get started, browse the preset connectors or learn more in the documentation. Read more

Vercel MCP now supports purchases

July 21, 2026

Vercel MCP now supports purchasing Vercel products. You can: Upgrade your team to the Pro plan Add prepaid credits for v0 (requires a paid v0 plan) or AI Gateway Purchase the SIEM add-on (requires an Enterprise plan) Purchase and register a domain Vercel MCP quotes the price, explains whether the charge is one-time or recurring, and completes the purchase only after you confirm. When an exact price isn't available, it links to the relevant pricing or billing page. Purchases require a team role with billing access and a valid payment method. Connect your AI assistant to Vercel MCP to get started. Read more

Laguna S 2.1 is now available on AI Gateway

July 21, 2026

Laguna S 2.1 from Poolside is now available on AI Gateway. There are 2 versions of the model available: Free version (256K context window): poolside/laguna-s-2.1-free Paid version (1M context window): poolside/laguna-s-2.1 Laguna S 2.1 is an open-weight Mixture-of-Experts model that supports a context window of up to 1M tokens and runs in thinking and no-thinking modes. The model specializes in agentic coding and long-running tasks, including writing and debugging code, running tests, building browser-based tooling, and working on MLOps pipelines and AI research. In thinking mode, Laguna S 2.1 reports 70.2% on Terminal-Bench 2.1, 78.5% on SWE-bench Multilingual, and 59.4% on SWE-Bench Pro. To use Laguna S 2.1, set model to poolside/laguna-s-2.1-free or poolside/laguna-s-2.1 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, includi

Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now available on AI Gateway

July 21, 2026

Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now available on AI Gateway. Gemini 3.6 Flash improves quality across coding, agentic tasks, and web development while consuming fewer tokens and making fewer model calls. It produces cleaner web and app development output. Gemini 3.5 Flash Lite upgrades the agentic capabilities of the Flash-Lite tier, making it a good fit for subagents that handle scoped parts of a larger task. To use them, set model to google/gemini-3.6-flash or google/gemini-3.5-flash-lite in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Gemini 3.6 Flash in the model playground. Read more

Service tiers now available on AI Gateway

July 21, 2026

AI Gateway now supports service tiering. Service tiers let you optimize for latency, throughput, and cost per request to match your use case. Pick a faster tier for interactive workloads (less queueing, higher token throughput), or a lower cost tier for background jobs that can tolerate more latency. At launch, service tiering is available for OpenAI and Gemini models. Service tiers work across every AI Gateway API format: AI SDK, Chat Completions API, Anthropic Messages API, OpenAI Responses API, and OpenResponses API. AI Gateway adjusts billing based on the tier each request used. Tiers default: Standard processing priority: Faster processing at increased cost flex: Lower cost with potentially higher latency If a service tier is not specified, requests run on the default tier. Basic usage Set serviceTier under providerOptions.gateway. The same option works across all models and providers that support service tiers, so you can swap the model without restructuring provider-specific options: The applied tier is returned in the provider metadata so you can confirm which tier served the request. This is useful when a priority request falls back to default capacity, or when comparing o

How Searchable ships customer-requested features in 30 minutes on Vercel

July 21, 2026

Searchable on Vercel 5x increase in development velocity 100+ billion tokens processed Customer-requested features shipped in as little as 30 minutes Zero model SDK implementation or API key rotation with AI Gateway Searchable helps brands track and improve how they appear across AI search engines like ChatGPT, Perplexity, and Claude, pairing visibility analytics with an agent that guides users on what to do next. The Searchable team builds on Vercel's AI SDK and AI Gateway to test new models without rotating API keys or SDKs. As a result, development now moves 2 to 5x faster. Read more

Introducing the new Vercel Agent

July 21, 2026

Today we're expanding Vercel Agent. It started by triaging alerts and reviewing your pull requests. Now it has a home in your dashboard, where it can investigate production, answer questions about your projects, and take action once you approve it. Because Vercel Agent is built into the platform that deploys and runs your app, when something changes in production, it's your first responder. It autonomously investigates your logs, metrics, and deployments, finds the root cause, and proposes a fix, before you've opened your laptop. Vercel Agent works under its own identity and is read-only by default. You can reach it through the Vercel Dashboard, GitHub, and the CLI. What Vercel Agent does for you We've been running Vercel Agent on our production deployments for months. This is what a typical investigation looks like. A bad deploy ships at 11pm and the checkout endpoint starts throwing 500s. By the time the on-call engineer logs in, Vercel Agent has already traced the errors to the deploy that shipped four minutes earlier, and recommends an instant rollback. The engineer approves the plan. With permission, Vercel Agent rolls back to the previous production deployment and begins work

Python function bundles now include precompiled bytecode

July 21, 2026

Vercel now compiles Python functions to bytecode at build time. In our benchmarks, cold starts for the median-sized function dropped from 2.8s to 1.3s. Cold starts are already rare in practice. Pre-warmed instances keep at least one function instance active for production deployments on paid plans, and automatic concurrency scaling reduces how often new instances start cold. Precompiling to bytecode makes starts faster across the board, so proactive initialization finishes sooner and fewer requests ever hit a cold start. When Python imports a module without cached bytecode, it parses and compiles the source before executing it. That compilation step adds startup time for functions with large dependency trees. Vercel now compiles both application code and dependencies and includes the resulting .pyc files in the function bundle, so the interpreter skips compilation at startup. Vercel automatically adds as much precompiled bytecode as fits in the function bundle. Functions near the size limit have little room left, so they see smaller gains. No code changes are required. See the Python documentation to learn more about Python support on Vercel. Read more

Configure where run state lives in Vercel Workflows

July 20, 2026

Vercel Workflows now keeps each run's state, queue dispatch, and output streams in a single home region: the region where the run starts by default, or any target region you choose. A run keeps its home region for its lifetime, so for agents built on Workflows, the whole loop stays near the user: an agent serving someone in Sydney executes, checkpoints its progress, and streams output from Sydney. During a regional incident, workflow traffic fails over to the next closest region. To get started, update the Workflow SDK to workflow@5.0.0-beta.33 or later: To pin a run to a specific region, pass the region option to start(): Existing workflows pick up regional placement on their next run, with no migration or code changes. Learn more in the Workflows documentation. Read more

Purge your Remote Cache's build and CI artifacts

July 20, 2026

Team Owners can now clear the team's Remote Cache of all artifacts in one click. This is useful when you believe there are poisoned artifacts in your cache. In your team's Build and Deployment settings, visit the Remote Caching section and clear the Remote Cache. Visit the docs to learn more. Read more

Data downloaded by Vercel Sandbox is now free

July 17, 2026

Vercel Sandbox no longer bills for data it downloads from the internet. Installing packages, cloning a Git repository, or pulling artifacts and datasets from an external source does not count toward Sandbox Data Transfer usage. Traffic received on a Sandbox's exposed ports is still billable, as is outbound traffic the Sandbox sends to the internet. Pricing for Active CPU, Provisioned Memory, Snapshot Storage, and Sandbox Creations is unchanged. See the pricing documentation for the full breakdown. Read more

Runtime logs now show cache reasons

July 17, 2026

Runtime logs now show a Cache Reason explaining why a request wasn't a fresh cache hit, for example a time-based or tag-based revalidation. Use cache reasons to debug misses and improve your hit rate. Cache reasons appear for any response the CDN can cache, including ISR, Partial Prerendering, and functions that set a Cache-Control header with directives like stale-while-revalidate. Responses rendered dynamically on every request don't have a cache reason. Cache status Possible reasons MISS Cold, Request collapsed, Error BYPASS Draft Mode, Prerender Bypass, Crawler STALE Time-based revalidation, Tag-based invalidation, Revalidation error REVALIDATED Tag-based deletion Open the Logs tab and select a request to see the reason alongside its cache status. Each status and reason links to its entry in the Cache Status and Reasons reference. From the CLI Cache reason is available in vercel logs, vercel metrics, and Vercel plugin's cdn-caching skill: Learn more about cache statuses and reasons. Read more

GLM 5.2 is 35% off via Novita on AI Gateway

July 17, 2026

GLM 5.2 is 35% off on AI Gateway through July 24 when routed through Novita. To get the discounted rate, set the model to zai/glm-5.2 in the AI SDK and route requests through Novita: After July 24, the model stays available at standard provider rates with no markup. Try GLM 5.2 in the model playground. Read more

Optimized CDN caching and deploying of immutable static assets

July 17, 2026

Vercel now reuses static files across deployments for frameworks that output content-addressed assets. It's zero-config: Vercel uses Framework-defined infrastructure to manage these immutable files alongside your code changes, handling the hard parts like hash collisions, file lifecycles, and efficient routing. This brings several benefits: Lower costs and improved cache hit rates — For frameworks using query-parameter-based Skew Protection (such as Next.js), we saw 17% fewer CDN requests and 24% fewer bytes transferred for static content. Faster Deployments — Deployments complete up to 30% faster on average. Faster Time to First Byte — Immutable assets can remain cached across deployments, giving frequently deployed projects up to a 60% reduction in global TTFB. Vercel's supported frameworks will automatically take advantage of this optimization, even for projects without Skew Protection enabled. Next.js 16.3 preview and later already have it enabled out of the box, and support for additional frameworks (including Nitro) is coming soon. No configuration is required: upgrading to a supported framework version is all you need to do. If you're a framework author and want to implement

Vercel Plugin now available in Kimi Code CLI

July 17, 2026

The Vercel Plugin is now available in the Kimi Code CLI. Kimi Code can now draw on Vercel platform knowledge on demand, with skills for Next.js, AI SDK, Vercel Functions, and more. The Vercel Plugin also helps Kimi Code stay up to date with the latest Vercel APIs and recommended patterns. To install it, upgrade to the latest Kimi Code CLI and select the Vercel Plugin from the Third-party section of the /plugin menu. Learn more in the Vercel Plugin documentation. Read more

Chat SDK adds native Slack agent support

July 17, 2026

You can now build native Slack agents with Chat SDK's Slack adapter. The adapter supports the full Slack agent messaging experience, from agent conversations in the Messages tab to suggested prompts, rotating status messages, token-by-token streamed replies, and native feedback buttons. Here's what the adapter gives you: Suggested prompts, per thread: Pass a static payload or an async resolver that receives the thread context, including what the user is currently viewing under agent_view. Prompts are pinned automatically whenever an agent thread opens. Native streaming with a fallback: Streamed replies render token-by-token via Slack's streaming API, including task and plan cards. If a workspace doesn't support streaming (e.g., GovSlack), the adapter switches to Post+Edit. Built-in feedback: Set feedbackButtons: true to append thumbs-up/down buttons to every streamed reply, and clicks dispatch via the bot.onAction flow. One thing to know: under agent_view, Slack threads each user message individually, so channel history only returns the user's side of a DM. Use Chat SDK transcripts to build AI conversation history instead. Read the documentation to get started, or begin with one of

Kimi K3 is now available on AI Gateway

July 16, 2026

Kimi K3 from Moonshot AI is now available on AI Gateway. K3 is an open-source model with a 1M-token context window and native visual understanding, accepting text, image, and video inputs. Built for long-horizon software engineering, knowledge work, and deep reasoning, K3 is especially strong where code meets visual and spatial reasoning, which suits frontend, game development, and CAD workflows. Thinking mode is always on. To use Kimi K3, set model to moonshotai/kimi-k3 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Kimi K3 in the model playground. Read more

Write utilization now available in ISR Observability

July 16, 2026

The ISR Observability page now shows Observability Plus subscribers a write utilization metric, helping identify routes that regenerate often but receive few requests. Write utilization is the ratio of cached requests to ISR writes. For routes with low write utilization, consider increasing the revalidation interval or switching to on-demand revalidation to reduce costs. You can also compute write utilization using the Vercel CLI, or use agents with the cdn-caching skill to investigate. Learn more about Incremental Static Regeneration or follow our guide on how to reduce ISR regenerations and costs. Read more

How Speechify serves 500,000 dynamic pages to 60 million users on Vercel

July 15, 2026

Speechify on Vercel 500,000+ pages served across 40+ languages Cut costs 50% by auto-scaling with Fluid compute Zero user impact on bad deploys with Instant Rollbacks Speechify started as a tool for people with dyslexia. Cliff Weitzman, Founder & CEO, built it because reading was challenging and audio made it much easier. That initial use case led Speechify to tens of millions of users and an Apple Design Award for inclusion. The product has since grown into something much larger: an AI work platform where 60 million people listen to documents, delegate tasks to agents, and complete work entirely through voice. But the journey wasn't easy. Early on, they got hacked, and for half a day, every visitor was redirected to a casino website. After that, Denis Chernobai, Head of Growth Engineering, knew it was time to re-evaluate their infrastructure stack. He decided to rebuild from scratch on Next.js and Vercel, and they ended up serving 40 times more pages, reaching a global audience three times larger, and achieving a 50% cost reduction. Since migrating to Vercel, Speechify has maintained 99.99% uptime and hasn't had a single security incident. The cost of serving 500,000 dynamic pages

Vercel Connect support in GitHub Tools

July 15, 2026

GitHub Tools now has first-class support for Vercel Connect through the new @github-tools/sdk/connect subpath. Instead of storing a long-lived personal access token, your agent mints short-lived, scoped GitHub tokens at runtime from a connector. There is no secret to store, rotate, or leak. Attach a GitHub connector to your project and pick a preset: Scopes from presets: Presets such as code-review, issue-triage, and maintainer map to Connect scopes automatically, so tokens carry only the permissions the toolset needs. Works with eve: Import from @github-tools/sdk/connect/eve and one file in agent/tools/ registers the full toolset with Connect-backed auth. Multi-tenant ready: Override installationId, repositories, or scopes per call to target a specific installation or narrow access to individual repos. Zero-config on Vercel: Deployments automatically authenticate with the OIDC token. For local development, run vercel link and vercel env pull. For custom tool factories, connectGithubToken returns a lazy token provider you can pass to createGithubTools directly. Get started by creating a GitHub connector and reading the documentation. Read more

Faster, predictable project linking in the Vercel CLI

July 15, 2026

The Vercel CLI now resolves your team before discovering projects, then searches for projects only in that team instead of sweeping every team you belong to. Linking is faster and more predictable, and every command that establishes a link (vercel link, deploy, pull, dev, and git connect) follows the same flow: Teams resolve from an explicit signal (such as --team, --scope, the scope field in vercel.json, or VERCEL_ORG_ID), if you only have one team, or from a searchable picker. Project suggestions begin with projects already linked to your local Git repository, then fall back to an exact folder-name match before opening up search and project creation across the team. The --yes flag answers confirmations without ever selecting a team for you. For CI and agent workflows, setting VERCEL_ORG_ID and VERCEL_PROJECT_ID makes vercel link fully promptless, while --team or --scope settles the team on any command. When more than one team is available and no signal is set, non-interactive commands fail with an action_required: missing_scope error (JSON output) before they create anything or touch an existing link. Vercel CLI 55 is a major version bump and includes breaking changes. If a scrip

Inkling from Thinking Machines is now available on AI Gateway

July 15, 2026

Inkling from Thinking Machines is now available on AI Gateway. Inkling is a broad generalist model, trained across agentic, reasoning, coding, instruction-following, factuality, vision, and audio tasks rather than optimized for a single domain. The model also supports controllable thinking effort. To use Inkling, set model to thinkingmachines/inkling in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Inkling in the model playground. Read more

Vercel Workflows trace viewer now has a minimap

July 15, 2026

The trace viewer for Vercel Workflows now has a minimap that shows the entire run in one view. Drag the viewport to pan, resize it to zoom, or drag across the minimap to select a new section. Click anywhere on the minimap to jump to that point, or scroll and use arrow keys to step through the run. The minimap is available in the trace viewer on Vercel and locally through the Workflow SDK with npx workflow@beta web. Learn more about Vercel Workflows and the Workflow SDK. Read more

Chat SDK adds Discord Components V2 support

July 15, 2026

You can now send Discord bot messages with Components V2, an opt-in layout system that treats text, images, files, and buttons as flexible components you can arrange in any order. Set the adapter's contentFormat option to ComponentsV2 and cards render with native containers, sections, media galleries, separators, buttons, and string selects. Embeds remain the default, so existing bots work unchanged. Individual sections can carry their own actions, markdown renders correctly inside components, and the adapter enforces Discord's platform limits for you, including the 40-component cap per message. This release also includes several other Discord adapter improvements: Rename threads: A new setThreadTitle()method renames native Discord thread channels. Your bot needs the Manage Threads permission to use it. Ignore global pings: Bots running in gateway mode no longer treat @everyone and @here announcements as mentions. A new respondToGlobalMentions config option (default false) lets you opt back in. Direct mentions, role mentions, and replies to the bot are unaffected. Precise mention detection: Mentions of similarly named users no longer trigger false matches. A message with @bot-dev n

Vercel Blob now supports consistent reads on private storage

July 14, 2026

Vercel Blob now supports consistent reads on private storage. Pass useCache: false to get() or presignUrl() for a read that reflects the latest write. A blob written to a fresh pathname has no existing cached entry, so reads reflect the latest write immediately. When you overwrite a blob at an existing pathname, readers might see the cached version for up to 60 seconds. When a read must reflect the latest write, like an agent's memory file, a session transcript, or a scheduled JSON report, pass useCache: false. These reads bypass the CDN, take longer than cached reads, and incur Fast Origin Transfer. You can also bypass the cache without the SDK. Under the hood, useCache: false adds a cache=0 query parameter to the request. You can set it yourself on a private blob URL: Install the latest @vercel/blob SDK for consistent reads: Learn more in the consistent reads documentation. Read more

Vercel Plugin now available in VS Code and GitHub Copilot CLI

July 14, 2026

The Vercel Plugin is now available in VS Code and the GitHub Copilot CLI. GitHub Copilot now has Vercel platform knowledge on demand, with skills for Next.js, AI SDK, Vercel Functions, and more. The Vercel plugin also helps Copilot stay up to date with the latest Vercel APIs and recommended patterns. To install it for Copilot Chat, search for @agentPlugins vercel in the Extensions panel in VS Code. To install it for the GitHub Copilot CLI, run npx plugins add vercel/vercel-plugin. Learn more in the Vercel Plugin documentation. Read more

Node.js 20 is being deprecated on October 1, 2026

July 14, 2026

Following the Node.js 20 end of life on April 30, 2026, we are deprecating Node.js 20 for Builds and Functions on October 1, 2026. How can I see which of my projects are affected? You can see which of your projects are affected by this deprecation with: Will my existing deployments be affected? No, existing deployments with Serverless Functions will not be affected. Invocations to already-deployed functions will continue to work normally. Only new deployments will be affected. When will I no longer be able to use Node.js 20? On October 1, 2026, Node.js 20 will be disabled in Project Settings. Existing projects using 20 as the version for Functions will display an error when a new deployment is created. How can I upgrade my Node.js version? You can configure your Node.js version in Project Settings or through the engines field in package.json. You can also hand the upgrade to your coding agent: What if I can't upgrade before October 1? If you can't complete the upgrade in time without disrupting your app, deploy it as a container image instead. Add a Dockerfile.vercel at your project root and the image is built and deployed on every commit: We still recommend upgrading when you can.

Endform joins the Vercel Marketplace

July 14, 2026

Endform is now available on the Vercel Marketplace. It runs your Playwright tests in parallel, so your suite finishes in the time of your slowest test. Key capabilities: Run every test on its own isolated machine Bring your existing Playwright tests with no config changes Get a pass/fail check on every production and preview deployment, linked to full traces in the Endform dashboard Spot flaky tests by tracking results across historical runs Pay only for the test runtime you use Install Endform on the Vercel Marketplace or with the Vercel CLI: vc i endform. Read more

AgentMail joins the Vercel Marketplace

July 14, 2026

You can now provision and manage AgentMail directly from the Vercel Marketplace. Your AI agents can send, receive, and respond to email from a real inbox. AgentMail handles the mail servers and deliverability. Key capabilities: Provision inboxes and custom domains through the API Handle threads, attachments, and deliverability out of the box Trigger agentic workflows using webhooks and WebSockets Install AgentMail on the Vercel Marketplace or with the Vercel CLI: vc i agentmail. Read more

Chat SDK adds X adapter support

July 14, 2026

Chat SDK now supports X, extending its single-codebase approach to Slack, Discord, GitHub, Teams, Telegram, and WhatsApp with the new X adapter. Teams can build bots that reply to public mentions and hold direct message conversations through the X API v2 and the X Activity API. The adapter handles CRC verification and webhook signature checks automatically and can manage OAuth token refresh for long-running bots. Likes are the only supported reaction, and responses post once on completion since X has no native streaming. X's automation rules apply to automated messages. Read the documentation to get started or browse the adapter directory. Read more

Access and share AI Gateway leaderboard data

July 14, 2026

We are making the data behind the AI Gateway leaderboards open under the CC BY 4.0 license. You can now download or query the data through the leaderboard-export API endpoint and render any chart as a shareable image. The AI Gateway leaderboards show how AI is used in production, ranking traffic for models, labs, apps, and providers. Data is aggregated daily across trillions of tokens, so you can see what gets adopted and how that changes over time. For deeper analysis, see the July AI Gateway Production Index. What's ranked There are four leaderboards, each with its own metrics: Leaderboard Ranks Metrics Models Individual models Requests, token volume, spend, images or videos generated Labs Model labs Requests, token volume, spend, images or videos generated Apps Opted-in apps built on the AI Gateway Token volume, spend Providers Inference providers Token volume, spend Models and labs can be filtered by modality (text, image, video) and show a daily percentage share over time; apps and providers are aggregated across all modalities and show a ranked top list. Open data The data behind the leaderboards is open, published under Creative Commons Attribution 4.0 (CC BY 4.0). You are f

Open-weight models surge to 29% of volume, price per token flattens

July 13, 2026

AI Gateway Production Index — July 2026 Every month, AI Gateway routes tens of trillions of tokens between production applications and AI labs, giving us a view of what AI usage actually looks like in today’s enterprise. We publish that view here. See the Production Index reports published in May and June. July 2026 summary The July index reports on AI Gateway data collected in June 2026. AI Gateway token volume grew 29% month over month and spend grew 27%. The price per token was flat after rising almost 20% in May. Open-weight models ran 29% of gateway tokens, up from 11% in April, on under 4% of spend. DeepSeek reached 22.6% of token volume, in third place and less than two points behind Google, and GLM 5.2 broke into the gateway's top models by volume two weeks after release. Anthropic took 61% of gateway spend on 32% of tokens, and captured 72% or more of spend in every high-stakes use case. Each modality has a different leader. OpenAI's GPT Image generated 53% of images, ahead of Google's Nano Banana at 39%. Chinese labs took two-thirds of video spend, while xAI's Grok Imagine generated 42% of videos, the most of any model. Claude Fable 5 was released June 9 and reached 22% o

Agent Runs now show subagent activity on eve projects

July 13, 2026

You can now inspect subagent activity for eve projects in Agent Runs. The new Subagents tab shows every subagent, organized by which turn started it. Each row shows the prompt, duration, and any failures, all on a shared timeline. Click any subagent to open its run. The tab shows the same details as a parent run: turns, tool calls, metadata, cost, and token usage. Open Agent Runs to try it, or read the documentation to learn more. Read more

Manage Vercel Flags targeting rules from the CLI

July 13, 2026

You can now manage targeting rules for Vercel Flags through the Vercel CLI. With the vercel flags rules command, you and your agents can add new rules, move existing ones, and inspect the current ordering without leaving your terminal. Rules you create from the CLI use the same model as the dashboard. Conditions can target entities or reusable segments, outcomes can serve a single variant, a weighted split, or a progressive rollout, and rules evaluate top to bottom. For scripts and agents, vercel flags rules ls --json prints the full rule set. An environment can inherit its flag configuration from another environment. When you add or update a rule in an inheriting environment, the CLI switches that environment to its own configuration, so your rule applies there without changing the environment it inherited from. Update to the latest Vercel CLI version and run vercel flags rules to get started. Learn more in the Vercel Flags CLI documentation. Read more

Configure which sources can create deployments with Deployment Policies

July 13, 2026

Deployment Policies are now available and enable teams to restrict which mechanisms, organizations, and repositories are allowed to create deployments. Each policy can be configured per environment at the team and project level, allowing flexible combinations of rules. Learn more about Deployment Policies. Read more

Web Analytics and Speed Insights are now more cost-efficient

July 13, 2026

Web Analytics and Speed Insights are now almost 10% more cost-efficient for all Pro and Enterprise teams. Infrastructure and event-processing improvements reduced cost across both products, so page views, referrers, top routes, and Core Web Vitals are now processed more efficiently, and savings scale with your event volume. The change applies automatically, with no configuration, plan changes, or code to update. You'll see the change reflected on your next bill. Learn more in the Web Analytics and Speed Insights documentation. Read more

Seedream 5.0 Pro is now available on AI Gateway

July 11, 2026

Seedream 5.0 Pro is now available on AI Gateway. Seedream 5.0 Pro is an image generation and editing model. It generates images from text, rendering text without spelling errors and following typographic rules, and produces dense infographics with charts, timelines, and layouts alongside realistic imagery. To use Seedream 5.0 Pro, set model to bytedance/seedream-5.0-pro in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Seedream 5.0 Pro in the model playground. Read more

Traces now support Tree and Waterfall views

July 10, 2026

Traces in Vercel Logs now support Tree and Waterfall views. See span hierarchy, critical path, and timing without leaving the log entry. The Tree view groups spans by parent and child. On load, spans are sorted by duration so the slowest appear first. The Waterfall view places every span on one chronological axis, showing what ran in parallel and where time was lost. Read the Tracing documentation to learn more. Read more

You can now deploy Lovable apps to Vercel

July 9, 2026

Vercel now supports Lovable applications with zero configuration. To get started, sync your Lovable project to GitHub, then import the repository to Vercel. Vercel automatically detects the framework (TanStack Start) and deploys your app. Once connected, every change you make in Lovable syncs to GitHub and triggers a new deployment on Vercel. Lovable projects now use Nitro under the hood. Nitro is the universal server toolkit that powers zero-config deployments for TanStack Start and many other frameworks on Vercel, so no manual build configuration is required. Learn more about deploying Lovable apps on Vercel. Read more

Build logs now redact Sensitive Environment Variable values

July 9, 2026

When an Environment Variable is marked Sensitive and has a value 32 characters or longer, Vercel now replaces it with [REDACTED] in the deployment's build logs. The Build Logs view also indicates when redaction happened, giving teams visibility into why output was masked without ever exposing the underlying value. When a value is redacted, Vercel logs the Environment Variable key, project, and deployment in the Activity Log, but never the value. Vercel also redacts selected system Environment Variables used for deployment security. Learn more about Sensitive Environment Variables. Read more

Muse Spark 1.1 is now available on AI Gateway

July 9, 2026

Muse Spark 1.1 from Meta is now available on AI Gateway. It is a multimodal reasoning model with a 1M token context window built for agentic tasks, accepting text, image, video, PDF, and audio inputs. Muse Spark 1.1 plans and orchestrates work across tools and services, operating as a main agent or as a subagent, and it works with new tools, MCP servers, and custom skills without examples. The model supports parallel tool calling, structured output, and built-in search with citations. To use Muse Spark 1.1, set model to meta/muse-spark-1.1 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Muse Spark 1.1 in the model playground. Read more

GPT 5.6 Sol, Luna, and Terra now available on AI Gateway

July 9, 2026

GPT 5.6 is now available on AI Gateway in three models: Sol, Terra, and Luna. GPT 5.6 from OpenAI is now available on AI Gateway in a limited preview, across three models: Sol, Terra, and Luna. All three are stronger at agentic work across coding, biology, and cybersecurity, and are more token-efficient than the previous generation. Sol (openai/gpt-5.6-sol): the flagship, and the most capable of the three. Terra (openai/gpt-5.6-terra): a balanced model for everyday work, with performance comparable to the previous generation at half the cost. Luna (openai/gpt-5.6-luna): a fast, affordable model with strong capability at the lowest cost in the series. To use GPT 5.6, set model to one of the above slugs in the AI SDK: You can also set routing rules to switch to GPT 5.6 from other gateway models without touching your code. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a pl

Flags SDK now evaluates flags 10x faster

July 8, 2026

Flags SDK and Vercel Flags now evaluate multiple feature flags in bulk around 10x faster. Flag evaluation time improved by reducing microtask queue overhead and creating fewer promises. The improvement scales with the number of flags. Use await evaluate([flagA, flagB]) instead of Promise.all([flagA(), flagB()]) to benefit from these optimizations: You can also pass an object to evaluate flags with named keys precompute() automatically benefits from these improvements as well. Upgrade flags and @flags-sdk/vercel to the latest versions to get started. Learn more in the documentation. Read more

Use any Chat SDK adapter with eve

July 8, 2026

eve now supports Chat SDK adapters with the new Chat SDK channel. One channel connects your eve agent to Facebook Messenger, WhatsApp, Resend, Liveblocks, and any other surface with an adapter. You write normal Chat SDK handler code, and calling send inside a handler hands the message to your agent. Register handlers on bot exactly as you would in a standalone Chat SDK app. Out of the box, the channel: Mounts a webhook route for each adapter (e.g., /eve/v1/resend) Shows a typing indicator while a turn runs, then posts the agent's reply Renders human-in-the-loop input requests as cards with buttons, and resumes the session when one is clicked Persists the thread, so later events, including proactive sends from schedules, reach the same conversation Reports failures in the thread with a readable message Supply your own events handlers to override any of these defaults. Read the Chat SDK channel documentation to get started, or learn how to build your own eve agent with Resend. Read more

Chat SDK now supports Vercel Connect

July 8, 2026

You can now use Vercel Connect to manage credentials for your Chat SDK bots, with no tokens or signing secrets to store or rotate yourself. The new @vercel/connect/chat subpath provides adapter helpers for Slack, GitHub, and Linear. Each helper takes a connector UID and returns a config you spread into the matching adapter factory. Outbound bot calls use a function-form token field backed by getToken, so each API request gets a fresh, short-lived token that Vercel Connect rotates for you. Inbound triggers use a webhookVerifier that validates the OIDC token Connect attaches to forwarded events, eliminating the need for a signingSecret. Read the documentation to set up your first connector with Chat SDK or follow the knowledge base guide to build your own Slackbot with Vercel Connect. Read more

Chat SDK adds Dial support

July 8, 2026

Chat SDK now supports Dial with the new vendor-official adapter. Build bots that send and receive SMS, MMS, and iMessage on a real phone number, with bidirectional media and inbound voice-call transcripts. Replies use the standard Chat SDK thread and message APIs, with HMAC-verified webhooks and stable per-conversation threading. Each phone conversation becomes a Chat SDK thread. Texts and media arrive as messages on that thread, and when a voice call ends, its transcript appears there too. Subscriptions, handlers, posts, and per-thread state work the same as with any other Chat SDK adapter. Read the Dial documentation to get started. Read more

Chat SDK adds Photon support

July 8, 2026

Chat SDK now supports Photon with the new vendor-official adapter. Build bots that send and receive iMessages directly and in group chats, share media, and react with native tapbacks. Run the adapter against Spectrum Cloud, your own server, or directly on a Mac. Each Photon conversation becomes a Chat SDK thread, and tapbacks come through as reactions. Handlers, posts, and subscriptions work the same as with any other adapter. Webhooks are HMAC-verified, and your bot can reply to a DM straight from a webhook delivery, no live connection needed. Read the Photon documentation to get started. Read more

Grok 4.5 now available on AI Gateway

July 8, 2026

Grok 4.5 from SpaceXAI is now available on AI Gateway. Built for coding, knowledge work, and STEM, the model accepts text and image inputs. Grok 4.5 supports low, medium, and high reasoning levels and defaults to high. Set the level with reasoning to balance speed against depth. To use Grok 4.5, set model to xai/grok-4.5 in the AI SDK: You can also set routing rules to switch to Grok 4.5 from other gateway models without touching your code. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, routing rules, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Grok 4.5 in the model playground. Read more

Update Project Settings from the Vercel CLI

July 8, 2026

You can now update a project's framework preset and build settings directly from the CLI, without opening the dashboard. Run vercel project update from a linked project, or pass a project name directly: Settings can also be returned to automatic detection instead of an explicit override: For automation or further inspection, return the result as JSON: JSON output includes whether anything changed, the list of changed settings, the project ID and name, and the requested settings. Output is written to stdout with no other output mixed in, so it's safe to pipe and parse. Agents can use this to repair misconfigured projects end to end: after diagnosing a failed build, an agent can correct the framework preset, build command, or output directory and redeploy, without dashboard access. Invalid framework slugs and settings are rejected with suggestions before any API call is made. Update the Vercel CLI to v54.21.1 or later with npm i -g vercel@latest to get started. Learn more in the Vercel project documentation. Read more

Vercel Microfrontends checks for missing configuration

July 8, 2026

The Microfrontends Config Present deployment check prevents broken production routing by blocking any default app deployment that is missing its microfrontends.json. This appears alongside other Native Deployment Checks, such as lint and typecheck. The check is blocking by default for eligible apps, with no opt-in required. Eligible apps are projects enrolled in Vercel Microfrontends and designated as the default app in a Microfrontends group. If microfrontends.json is missing from the deployment's build outputs, the check fails and prevents the deployment from reaching production. This can happen when a deployment from an older branch without the Microfrontends config is manually promoted. The check passes on the next deployment once microfrontends.json is included in the build outputs. To adjust check behavior, visit your project's Build and Deployment Settings. Read more

More granular observability for Vercel Sandbox

July 7, 2026

Vercel Sandbox observability now includes detailed resource metrics, giving you deeper visibility into how your sandboxes consume compute and networking. From the Observability tab in your dashboard, you can now monitor: Active CPU and CPU usage: Measures when your code actively uses the CPU. Active CPU is measured in core-hours and time spent waiting on I/O, such as network requests or model calls, is not billed Provisioned Memory: Memory allocated to your sandboxes multiplied by runtime, measured in GB-hours Data Transfer: Total data transferred in and out of your sandboxes, including package downloads and API calls Running sandboxes and sessions: How many sandboxes are running at any give time, and when sessions are stopped or started Each metric can be grouped by Sandbox Name and Sandbox Session ID, so you can drill down from aggregate usage to the individual sandbox responsible. You can also query and visualize metrics via the Vercel CLI: Metrics are available at both the team and project level and align directly with how Sandbox usage is billed, so you can attribute costs to specific workloads and catch unexpected usage early. This is useful for tracking agent workloads that

Give your eve agent GitHub tools

July 7, 2026

GitHub Tools now ships an eve toolset through the new @github-tools/sdk/eve subpath. One file in agent/tools/ can register every GitHub tool, or use a preset such as maintainer, so you can build a complete GitHub agent in nine lines of code. Safe by default: Every write tool, such as mergePullRequest, requires approval unless you opt out. Gate individual tools with always, once, or an input-dependent predicate; pauses survive restarts and deploys. Presets: code-review, issue-triage, repo-explorer, ci-ops, and maintainer scope the toolset, alone or merged. Trimmed reads: High-volume read tools such as listPullRequestFiles and getCommit trim what the model sees, while channels still get full payloads. Get started by following the knowledge base guide. Read more

You can now view the Activity Log at a project-level

July 7, 2026

Project Settings now includes its own project-specific Activity Log. This view pulls the relevant user events from the Activity Log in Team Settings. Review project-level activity without leaving your project settings. Visit your project settings to try it out. Read more

Vercel acquires Better Auth to accelerate open source auth

July 7, 2026

Vercel is acquiring Better Auth, the company behind the open source TypeScript authentication library, which has 4.7M+ weekly npm downloads and more than 850 contributors. Founder Bereket Engida and the core team are joining Vercel to continue their work on Better Auth and agent identity. The open web Last year, we published our approach behind Next.js, AI SDK, and Nuxt: software that's open by default, loosely coupled, and portable to any platform. Better Auth applies it to authentication: framework-agnostic, runs anywhere, and gives developers auth they own. Agent identity When an agent acts on your behalf, it runs under your identity and access, so every service it touches sees you, not the agent. There's no clean way to limit what any one agent or subagent can do, or to shut down just one without cutting off the rest. The Better Auth team has been building Agent Auth so that each agent can carry its own identity and its own scoped, revocable authority, with you as the single point of control. Agent identity is foundational to agentic infrastructure. The Better Auth team will continue that work at Vercel, bringing agent identity to Vercel Connect and eve. What's next for Better

Vercel Sandbox now supports FUSE-based filesystems

July 3, 2026

Vercel Sandbox now supports FUSE, letting you mount remote storage and custom filesystems inside a running Sandbox. Use it to attach S3 buckets, network filesystems, or any other FUSE-compatible driver as a regular path. This makes it possible to stream large datasets directly from object storage, share state across Sandboxes through a common filesystem, or run tools that expect POSIX paths against remote sources without copying data into the Sandbox first. Learn more about remote storage mounts in the documentation. Read more

Manage Vercel Flags segments with Vercel CLI

July 3, 2026

Vercel Flags segments can now be managed from the Vercel CLI with the new vercel flags segments command. A segment is the targeting primitive a flag uses to decide who sees what. Membership composes from three repeatable tokens: include:, exclude:, and rule:. Pass them to --add or --remove for incremental edits. For full replacement, --data takes the entire segment definition as raw JSON. All segment commands support --json output, making them scriptable from CI, local workflows, and agent-driven pipelines that need to inspect or update flag targeting from the terminal. Update to the latest version of the Vercel CLI and learn more in the Vercel Flags CLI documentation to get started. Read more

Agent Runs now available in the Vercel MCP and CLI

July 3, 2026

Your agent can now inspect Agent Runs via the Vercel MCP and CLI for eve, the open-source agent framework. eve traces are automatically ingested when deployed to Vercel and available as Agent Runs. The new Vercel MCP tools and Vercel CLI commands let you find projects with runs, list recent runs, and retrieve full traces, including reasoning, tool calls, and token usage. Vercel MCP tools: list_agent_run_projects: Find projects in a team with Agent Runs activity. list_agent_runs: List recent runs for a project. get_agent_run: Inspect metadata, lifecycle events, usage, and subagent data. get_agent_run_trace: Retrieve trace data for a run, including turns, messages, reasoning, tool calls, token usage, and tool input/output. Vercel CLI commands: vercel agent-runs projects vercel agent-runs list vercel agent-runs inspect vercel agent-runs trace Every CLI subcommand supports --json for machine-readable output, and traces render as markdown when piped, so coding agents without MCP access can call the CLI directly to debug their own runs. Ask your coding agent questions like "Show me the latest production Agent Runs for my project" or "Update skills based on recent runs". Get started by in

Routing rules now available on AI Gateway

July 2, 2026

Vercel AI Gateway now supports routing rules. Routing rules are firewall-style rules that control which models your team can use, applied at the gateway level instead of in your application code. When a model goes down or gets retired, you usually have to ship a code change to move off it. With routing rules, you push one rule and every request reroutes instantly. There are two types: Type What it does Use it to Rewrite Serves a request for one model using another Keep traffic flowing when a model is unavailable, migrate off a retired model, standardize on one model, or route an expensive model to a cheaper one Deny Blocks requests for a model Keep your team off models you haven’t approved Rules apply to every request made with your team's AI Gateway credentials. You manage them with the Vercel CLI. Rewrite Create a rewrite with a source and a destination model. The Gateway swaps in the destination transparently, so your application keeps requesting the source model: Deny Create a deny rule to block a model. Requests for it return a 403: Rules only change which model serves a request. Everything else you’ve configured still applies to the destination model, including: Request-level

Claude Fable 5 access restored on AI Gateway

July 1, 2026

Access to Claude Fable 5, the Mythos-class model, has now been restored on AI Gateway following the US Government's decision to lift the export controls. Fable 5 is the same model that was available between June 9 and June 12. What has changed is the safety classifiers, which are now updated and more robust. In the near term, some routine tasks such as coding and debugging may trigger safety classifiers. To ensure requests are still serviced when the safety classifiers are triggered, use model fallbacks. AI Gateway will try each model in models in the stated order if Anthropic refuses the request to Fable 5. To call Fable 5, use model name anthropic/claude-fable-5: Model fallbacks work on every API format: for more information on how to configure these, see the docs. Anthropic does not support Zero Data Retention for the model, because some misuse patterns are only visible across cumulative requests, which real-time filters cannot catch on their own. Prompts and completions are retained for 30 days and are not used to train Claude. Read more in the data retention whitepaper. Read more

Enforce consistent code for agents and humans with konsistent

July 1, 2026

konsistent is now open source. konsistent is a CLI linter for TypeScript codebases that enforces structural conventions, giving agents and humans the consistent context they need to implement features correctly. Deterministic, fast, and covers structural patterns that TypeScript and ESLint don't model. Configure conventions in a project-level konsistent.json file for concerns like: Do all files matching pattern X export functions Y and Z? Does every folder that has file X also have file Y? Does every class exported from files matching pattern X implement type Y? konsistent is used in AI SDK and Chat SDK to enforce structural code conventions. Use the konsistent skill to help your agent create a konsistent.json: Read more

Dry-run deployments with Vercel CLI

July 1, 2026

You can now preview the framework preset and files that Vercel CLI includes in a deployment before creating one. Run vercel deploy --dry from a linked project: For automation or further inspection, return the complete file manifest as JSON: JSON output includes the detected framework, included and ignored paths, directory size distribution, largest files, file modes, and content hashes. Piped and other non-TTY output automatically uses JSON. Agents can use this manifest as a pre-deployment check, verifying framework detection, identifying unexpected or missing files, flagging oversized assets or unusual file modes, and updating .vercelignore or project configuration. Agents can rerun the check until the manifest matches the intended deployment, without uploading code or creating a deployment. Update the Vercel CLI to v54.17.2 or later to get started. Learn more in the vercel deploy documentation. Read more

Vercel Security Dashboard is in private beta

July 1, 2026

The Vercel Security Dashboard is now in private beta. It aggregates the security posture of every account and project on Vercel. As teams grow and coding agents make it easy to spin up new projects, small misconfigurations can add up quietly and quickly. Team members without 2FA, publicly accessible preview environments, or long-lived credentials where short-lived ones would do. Each of these can lead to a security incident or breach. The security dashboard flags these findings across Vercel environments, explains what they are, and guides you towards fixing them. Sign up here to join the waitlist. Read more

Resend joins the Vercel Marketplace

July 1, 2026

You can now provision and manage Resend directly from the Vercel Marketplace. This integration helps developers start sending messages within minutes, build emails as React components, and track delivery events in real time. Key capabilities include: Send transactional and marketing emails through an API or SMTP relay Build email templates as React components with React Email Track opens, clicks, bounces, and deliveries with real-time webhooks Use the Chat SDK adapter or eve channel to send emails with your agent Install Resend on the Vercel Marketplace or with the Vercel CLI: Read more

Secure internal communication between services (beta)

July 1, 2026

Service Bindings make it easy for one Vercel service to securely call another within the same deployment. When a service declares a binding for another service, Vercel automatically injects the configured environment variable. That means user code can fetch that URL normally, while Vercel handles the internal rewrite, routing, authentication, and TLS behind the scenes. This enables multi-service applications on Vercel, like a Next.js frontend calling a FastAPI backend, while preserving service isolation and keeping routing configuration simple: In the example above, my_frontend will be able to reach the my_backend service by making a fetch() request to BACKEND_INTERNAL_URL. On the FastAPI side, my_backend exposes a normal route: Under the hood That URL points inside Vercel, not at a public address. When your code fetches it, the request travels over Vercel's internal network to the same routing layer that handles every other request to your deployment, which sends it to the target service. It never hits the public route table, and Vercel sets up the TLS trust for you, so a plain HTTPS fetch works with no certificate config. A service is only reachable if you expose it, either publi

Vercel Sandbox now support Custom Images

June 30, 2026

Vercel Sandboxes now supports custom images. Launching in public beta today, images allow Sandboxes to start with your own custom root filesystem. Images are pulled from Vercel Container Registry, so anything you docker push is immediately available. Bring your own OS, toolchain and dependencies into the Sandbox without needing to spin up compute and create Snapshots. Images in the background for Fluid Compute and boot from a precompiled snapshot in the same format as our Sandbox Snapshots - so you get the convenience of a custom filesystem without sacrificing cold start performance. Read more about running images with Sandbox in the documentation. Read more

Introducing VCR: Vercel Container Registry

June 30, 2026

You can now push, pull, and manage container images directly on Vercel. Vercel Container Registry is an OCI-compliant image registry hosted on Vercel infrastructure. It works with standard workflows like docker push, docker pull, and docker tag, so there's nothing new to learn or tooling to migrate. A Vercel project can have unlimited repositories. You can create and manage them through the Vercel dashboard, or push directly to VCR, which creates the repository for you on the fly. Operations use the same authorization controls as the rest of Vercel. You can authenticate with OIDC or an access token, provided the token has access to the project scope. Optimized for Fluid Compute When you push an image, VCR automatically optimizes it in the background for use in Sandboxes and Functions. VCR stores a precompiled snapshot and serves that snapshot when the image is requested. These snapshots use the same format as Sandbox Snapshots, which are optimized to run on Fluid Compute. Build and deploy on Vercel The Vercel build environment is automatically set up with authentication for the current project, so you can build and store images as part of the build and deploy process. You can also

Run multiple frameworks in one project with Vercel Services

June 30, 2026

You can now deploy multiple frontends and backends together within a single Vercel project. Vercel Services is now available, allowing you to deploy full stack apps with multiple frameworks on a shared domain, where services talk to each other privately and deployments build, preview, and roll back together. Services are defined in vercel.json: From there, your services show up across the dashboard and CLI: The Deployments panel visualizes the services graph The Logs UI filters by individual service vercel dev runs every service locally for a production-like environment Service bindings Services talk to each other internally with the new bindings key, without routing through the public internet: The frontend reaches the backend privately through the URL in BACKEND_INTERNAL_URL: Framework-defined infrastructure Most frameworks run with zero configuration. Framework-defined infrastructure means each service's framework is auto-detected and auto-provisioned, from FastAPI and Flask to Express and Hono, with first-class support for Go and Rust. Services run on Fluid compute with Active CPU pricing, so you only pay for the time your code is actually running. Read the documentation to get

Bring your Dockerfile to Vercel Functions

June 30, 2026

Vercel Functions now support deploying HTTP servers from a Dockerfile or Containerfile, using Open Container Initiative (OCI) compatible images on Fluid compute. This makes it easier to bring existing applications written in any language to Vercel while keeping the preview deployments, logs, routing, and autoscaling benefits of using Vercel Functions and Fluid compute. To use container images with your functions, create a project with a Dockerfile.vercel (or Containerfile.vercel) file that starts an HTTP server listening on $PORT: This image will be built, pushed, and deployed on every commit to Vercel Container Registry. Learn more about using container images in the docs. Read more

Expanded Audit Log coverage, now delivered through Vercel Drains

June 30, 2026

Audit Logs now capture 400+ unique team activity events, giving teams broader coverage for security reviews, compliance workflows, and investigations. With Vercel Drains support, teams can export those events to custom HTTP endpoints or Amazon S3, replacing Custom SIEM Log Streaming with a flexible Drains-based workflow. Audit Log Drains are available for all Enterprise teams, replacing Custom SIEM Log Streaming. They are priced at $0.50/GB through standard Drains pricing, replacing a separate paid add-on. Try it out or learn more about Audit Log Drains and draining Audit Logs to S3. Read more

Vercel and Shopify are rebuilding Hydrogen

June 30, 2026

Hydrogen made headless storefronts easy to ship, but not portable. At Vercel Ship 26 in New York, we announced that we are working with Shopify to rebuild it from the ground up, a shared bet on a more open web. The new version is open source and runtime agnostic, meaning it can run anywhere JavaScript does. You can build with Svelte, Nuxt, Next.js or even bring your own custom framework. Our strategy includes three layers: core, client and server. The core Core is the JavaScript we all used to write for the Shopify API, and never shared. Now it lives in one place. Take formatMoney. The open web already solved most of this with Intl.NumberFormat. But the Shopify API doesn't hand you a number. It responds with a custom type, MoneyV2, and the amount is a signed decimal number serialized as a string. The result is the same, but you’re not writing or maintaining the glue code anymore. When the API changes, the upgrade is trivial. Centralize the core and you fix each bug once, ship improvements to everyone, and get back to building. The client Rendering what the core returns involves the same repeated decisions. Cart state is the obvious one. Anyone who's built a commerce app has written

Vercel Services: Run full stack on Vercel

June 30, 2026

To a user, an app with a Next.js frontend and a FastAPI backend feels like one product, and the same should be true for the engineers who build it. Instead, the two pieces are often deployed across different clouds with different development and deployment workflows. Today we're introducing Vercel Services in public beta, which lets you run multiple frameworks in one Vercel Project. This unlocks: Atomic deployments: Your frontend, backend, and other services stay in sync and deploy or roll back together Shared preview deployments: See how any change affects all your services Internal service communication: Services can talk to each other without routing through the public Internet Vercel handles the rest: routing, builds, deployments, and auto-scaling in production. The developer experience you already know from Vercel now covers your entire application. Compose applications with Vercel Services Declare your services under the services key in vercel.json keeping the routing configuration explicit: Services can be mounted to a shared routing table without the need for a reverse proxy or CORS. The services configuration is recognized at multiple levels of the Vercel platform: The Dep

Vercel Agent has updated pricing

June 30, 2026

Vercel Agent pricing is changing. You no longer need to pre-load credits or manage a separate wallet for Vercel Agent. Instead of a $0.30 per-request fee, you now pay a Vercel Token Rate of $0.25 per million tokens, plus provider inference costs at the underlying token rate. The new Vercel Token Rate scales with the actual work each task requires. A quick question costs less than a deep investigation that reads your logs, deployments, configuration, and runtime data, spins up sandboxes, and writes across projects. You pay for the intensity of the work, not a flat fee regardless of what Agent did. The Vercel Token Rate covers: Joining your project context: logs, deployments, configuration, and runtime data Custom model routing and execution across your projects Processing and infrastructure costs It applies to input, output, and cached tokens, and is charged in addition to provider inference costs. For new Vercel Agent users, the Vercel Token Rate begins applying today. If you're already using Vercel Agent for code reviews, nothing changes for the next 30 days. You'll continue on per-request pricing for code reviews through that window, then automatically move to the new pricing, no

Claude Sonnet 5 now available on Vercel AI Gateway

June 30, 2026

Claude Sonnet 5 from Anthropic is now available on AI Gateway. Sonnet 5 improves on Sonnet 4.6 across coding and agentic work, reaching outcomes on many tasks that previously needed an Opus model, at Sonnet pricing. The model is more agentic and follows instructions more closely. Document parsing and long-context memory use are also stronger. Sonnet 5 also uses an updated tokenizer, like the recent Opus models, which can map the same input to more tokens. Launch pricing of $2 per million input tokens and $10 per million output tokens runs through August 31, 2026. Standard list price will be $3/M input tokens, $15/M output tokens. To use Sonnet 5, set model to anthropic/claude-sonnet-5 in the AI SDK: You can also try Sonnet 5 in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Read more

Vercel Private Blob is now generally available

June 30, 2026

Vercel Private Blob is now generally available for all plans. Store sensitive files like user-uploaded photos, invoices, and agent memory, and control exactly who can read them. Private stores, Signed URLs, and OIDC authentication all graduate from beta with this release. Vercel Private Blob uses the same Blob API as public blobs. Specify access: 'private' when uploading a blob: OIDC authentication Functions running on Vercel now authenticate to Vercel Private Blob with a short-lived, auto-rotating OIDC token scoped to the project, with no static read-write token in your environment. The Vercel CLI also supports OIDC authentication, so you can read from and write to a private store from your terminal without a long-lived token. For stores still using a static token, you can upgrade to OIDC and revoke the old credential from the dashboard. Signed URLs Mint a URL scoped to a single operation, pathname, and an expiration date you choose for up to 7 days, then hand it to a client to perform a PUT, GET, HEAD, or DELETE operation on a single object, without ever exposing your store credentials. Signed URLs work well when you want to grant temporary, narrow access to a private object with

An expanded Vercel Agent: chat, investigations, and approved actions, now in public beta

June 30, 2026

Today, we're launching expanded capabilities for Vercel Agent in public beta. Vercel Agent now lives in your dashboard and can investigate production issues, answer questions about your projects, and take action on your behalf. Because Agent runs inside the platform that deploys and serves your app, it has access to the signals around it: deployments, logs, metrics, project configuration, usage, and connected repositories. That context is what turns a question into an answer and a problem into a fix. What's new in this beta: Dashboard chat: ask questions about your projects and get answers grounded in your real deployment data, right from the dashboard. Production investigations: point Agent at a failed deployment, a runtime error, or a cost spike, and it traces the cause and recommends a fix. Approved actions: with your sign-off, Agent can open a PR, roll back, or update a config to remediate an issue. Agent code review works as it does today: Sandbox-validated suggestions on your pull requests, now one capability within a broader Agent. You stay in control Vercel Agent runs under its own identity and is bounded by the requesting user's permissions. It is read-only by default. Whe

Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) now on AI Gateway

June 30, 2026

Nano Banana 2 Lite from Google is now available on AI Gateway. This Flash-Lite-tier image model is built for fast, low-cost generation. It generates images alongside text in <4s and can edit existing images across multiple turns. The cost is also lower than previous Nano Banana models. Nano Banana 2 Lite generates 1K images at $0.034 each, about half the cost of Nano Banana 2 and roughly a quarter of the cost of Nano Banana Pro at the same resolution. This model is multimodal. Use streamText or generateText to generate images alongside text responses. To use Nano Banana 2 Lite, set model to google/gemini-3.1-flash-lite-image in the AI SDK: Here is the example output from the above prompt: You can also try Nano Banana 2 Lite in the model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Read more

Run any Dockerfile on Vercel

June 30, 2026

You have a server in a container. Maybe it's a Go service, a Rails app, a Spring Boot API, or a web server behind nginx. It speaks HTTP. It listens on a port. It just needs somewhere to run. Add a Dockerfile.vercel file to your project, and Vercel builds, stores, deploys, and autoscales the image on Fluid compute, so you pay only for the CPU your code uses. No daemon to run locally, registry to set up, or cluster to babysit. How it works Here is a small HTTP server in Go, listening on $PORT: Add a Dockerfile.vercel file that builds it into a small image and runs it: Then deploy: That is it. Two files, and you are live. Every git push rebuilds the image and hands you a fresh preview URL. Or run vercel to deploy without committing. We used Go in this example, but any language works. Rails, Spring Boot, Express, Laravel, ASP.NET, FastAPI, and a web server behind nginx all deploy the same way. The only rule is that your server listens on $PORT, which defaults to 80. If it speaks HTTP, it deploys. Yes, even Java. And yes, even PHP. What you get A container on Vercel is a first-class citizen. It runs on the same platform, and the same compute, as your frontend and the rest of your servic

Vercel Ship 2026 recap

June 30, 2026

For a decade, Vercel has shaped how the web is built. Now, we’re doing the same for agents. The companies that win the next decade will build on infrastructure designed for agents from the start, and thousands of people gathered across London, Berlin, and New York to do just that at Vercel Ship 2026. Agentic infrastructure Guillermo kicked off Ship by sharing his vision for Vercel: a true full-stack platform where you can deploy anything, including software that can think. Agentic infrastructure has three parts: Vercel is where coding agents deploy software. When you ask Claude Code or Codex where to deploy, you get Vercel, because Vercel is built for the way agents work. Vercel is where you build and deploy your own agents. We give you every tool you need to build and run apps and agents in production, securely, at scale. Vercel itself is automated by agents. Vercel runs your apps in production, handling traffic, traces, observability, and anomalies. That data gives our agents the context they need to investigate autonomously, and then surface pull requests, not just alerts. Vercel is for shipping agents In the next keynote, Tom walked through each primitive in the Agent Stack, an

Query Speed Insights from the Vercel CLI

June 29, 2026

You can now query Speed Insights datapoints directly through the Vercel CLI. Using the vercel metrics command, you can pull core Web Vitals (LCP, INP, CLS) and other page performance metrics (FCP, TTFB) based on client-side measurements from real user traffic. By providing a coding agent access to the CLI, an agent can answer questions such as: Which pages' INP have regressed since last week? How is the perceived speed of my home page in Asia? Compare the CLS of the dashboard between mobile and desktop. For more details on how to use the CLI command and a complete list of supported metrics, dimensions, filters, and query options, explore the documentation. Read more

Build realtime voice agents on AI Gateway

June 29, 2026

AI Gateway now supports audio/voice. You can add realtime voice, text to speech, and speech to text with the same calls you already use for text, image, and video, routed through AI Gateway alongside every other modality. Audio launches with models from OpenAI and xAI. Each call gets the same provider routing, observability, spend controls, and bring-your-own-key support you already use for your other models. These capabilities are in beta and available in AI SDK 7. Capability How it works Use it for Realtime voice Live audio in and out, for streaming, low-latency session Two-way voice agents and live conversation Text to speech Text in, audio file out, single request Voiceovers, spoken responses, audio versions of written content Speech to text Recorded audio in, text out, single request Transcribing voice notes, call recordings Getting started Realtime, speech, and transcription model are supported on AI SDK 7. Realtime voice agents Realtime turns your app into something a user can hold a conversation with. When they speak, the model responds right away. Because it replies in the moment instead of waiting for a full turn, users can interrupt and talk over it the way they would wi

Sandboxes now expire based on last use

June 29, 2026

Vercel Sandbox snapshots now expire based on when they were last used, not when they were created. Active snapshots stay alive as long as workflows depend on them, while unused snapshots expire on their retention policy. Every time a snapshot is used, its expiration timer resets. This lets you set shorter retention windows without worrying that a snapshot will disappear between sessions, making it safer to build long-running workflows on top of Sandbox persistence. Learn more about Sandbox snapshots in the documentation. Read more

xAI Grok audio models now available on Vercel AI Gateway

June 29, 2026

xAI's audio models are now live on AI Gateway. Realtime voice, text to speech, and speech to text are all available through the AI SDK with the same routing, observability, and spend controls as your other models. These capabilities are available on the AI SDK 7 release. Available models Capability Models Realtime voice xai/grok-voice-think-fast-1.0 Text to speech xai/grok-tts Speech to text xai/grok-stt Realtime A voice agent has two pieces: a server route that mints a short-lived token, so your API key never reaches the client, and a browser component that connects with it. Add the token route: this example sets model to xai/grok-voice-think-fast-1.0: Then connect from the browser. The useRealtimehook from @ai-sdk/react fetches that route and manages the WebSocket connection, microphone capture, and audio playback: Text to speech Generate spoken audio from text with generateSpeech. Pass a voice and an output format, then write the result to a file with xai/grok-tts: Speech to text Transcribe recordings into text with transcribe. This example uses xai/grok-stt: Playground You can also try the xAI audio models directly in the AI Gateway playground. Open the models list and click in

Realtime voice, speech, and transcription now supported on AI Gateway

June 29, 2026

AI Gateway now supports voice and audio models. You can build realtime voice agents, generate speech from text, and transcribe audio to text. This provides the same observability, spend controls, and bring-your-own-key support as text, image, and video models in AI Gateway, with no markup or platform fees. These capabilities are in beta and available via AI SDK 7. With realtime support, a single model takes audio in and audio out, so a user can talk and hear a reply back in near real time instead of waiting on a chain of separate models. Capability What it does Realtime voice agents Model listens to the user, works out a response, and speaks it back in a live, low-latency conversation. It can call your tools mid-conversation to look something up or take an action. The useRealtime hook handles microphone capture and playback. Text to speech Generate spoken audio from text, with a selectable voice and output format such as MP3. Use it for voiceovers, audio versions of written content, and spoken responses. Speech to text Transcribe recordings into text, from a file buffer, base64 string, or URL. Use it for voice notes or other transcriptions. Two ways to get started: Follow the realt

Vercel Open Source Program: Spring 2026 cohort

June 29, 2026

The world runs on open source software. The frameworks, libraries, and tools we rely on are made possible by communities that share ideas and build in the open. At Vercel, we want to help those communities thrive. That’s why we run the Vercel Open Source Program: a developer initiative that gives maintainers the resources and support they need to keep shipping and growing their projects. Four times a year, we welcome a new cohort of open source projects into the program. Maintainers receive $3,600 in Vercel platform credits, a starter pack with credits from third-party services, and guidance from Vercel and the community. Today, we welcome a new cohort into the program. Spring 2026 cohort ASCII Studio A browser-based tool that converts video frames in real time into precise ASCII representations using an optimized rendering pipeline. Users can adjust density, contrast, and character sets to produce customized ASCII animations. Creator: Vansh Nagar View the project View GitHub repository Sync UI A free, open-source React UI library offering 125+ animated components, 13+ pre-built blocks, and 3 premium templates. Built with Material UI and Motion (motion/react), designed for design e

Vercel Functions can now be up to 5GB in package size

June 29, 2026

Vercel Functions now support Node.js and Python deployments up to 5GB in package size on Fluid compute, a 20x increase over the previous 250MB limit. Large Functions support is in public beta. This makes Vercel compatible with backend workloads that do not fit within the 250MB limit, such as Python data and AI libraries, large generated clients, browser automation dependencies, image and video processing packages, and routes with substantial shared application code. New projects are automatically enrolled in the beta. Existing projects can opt in by adding VERCEL_SUPPORT_LARGE_FUNCTIONS=1 as an Environment Variable, then redeploying, or you'll be asked to opt in if a Function ever crosses the 250MB limit. You can scope the variable to preview deployments first to test before enabling it in production. Vercel still keeps Functions under 250MB in package size on the standard path. In eligible projects, only Functions that exceed the standard 250MB limit use the beta; other Functions in the same deployment are unchanged. The dashboard surfaces when a deployment is using Large Functions, so you always know exactly what's running. Learn more in the Vercel Functions limits documentation.

Query Web Analytics from the Vercel CLI

June 26, 2026

You can now query Web Analytics datapoints directly through the Vercel CLI. Using the vercel metrics command, you can pull page views, visitors, and custom events for your Vercel projects to analyze traffic, compare trends, and answer questions about site performance. By providing a coding agent access to the CLI, an agent can answer questions such as: Which pages gained the most traffic this week? Which UTM campaigns drove the most signups this month? Compare conversion events between mobile and desktop users. For more details on how to use the CLI command and a complete list of supported metrics, dimensions, filters, and query options, explore the documentation. Read more

Trace and debug eve agent sessions with Vercel Observability

June 26, 2026

You can now view Agent Runs in your Vercel dashboard for eve, the open-source agent framework. The Agent Runs tab appears automatically for every eve project, surfacing trigger, duration, and token usage for each session at a glance. Drill into any run to inspect every turn, model call, and tool call in the conversation. Runtime errors that previously vanished into function logs now correlate to the failing step. Two views read the same data: Developer mode shows raw tool names, input and output JSON, and per-step token counts for engineers debugging a failure. Business mode humanizes tool names, hides JSON, and generates a plain-English summary so a non-technical reviewer can audit what an agent did. Run data is encrypted by default. Retention scales with plan: Hobby: 12 hours Pro: 1 day Enterprise: 3 days Pro and Enterprise teams can extend retention to 30 days with Observability Plus. Teams using a custom OpenTelemetry backend can still export AI SDK spans from agent/instrumentation.ts to any destination, such as Braintrust or Datadog. Open Agent Runs in your project and read the documentation to get started. Read more

AI SDK 7

June 25, 2026

AI SDK, with over 16 million weekly downloads, is the TypeScript SDK for building AI applications, features, frameworks, and agents across any model provider. It's the same layer eve, Vercel's open-source agent framework, is built on. AI SDK 7 adds production depth for agent work across five areas: Develop agents with reasoning control, tool and runtime context, provider files and skills support, MCP Apps, and a terminal UI. Run agents with tool approvals, durability (WorkflowAgent), timeouts, and sandbox support. Integrate any agent harness, such as Codex, Claude Code, Deep Agents, OpenCode, or Pi. Observe agents with telemetry, Node.js tracing channel, lifecycle events, and performance statistics. Go beyond text agents with provider-agnostic real-time voice support and video generation. Develop agents Building well-behaved agents requires fine-grained control over model reasoning, tool context, and file handling. Reasoning control Most frontier models support configurable reasoning, but every provider API exposes it differently. AI SDK 7 standardizes this with a reasoning option for generateText and streamText. It maps to provider-native reasoning settings, letting you control re

Teaching agents product design at Vercel

June 25, 2026

Coding agents can produce working UI fast, but what's harder is a different shape. They can copy your product's style, match its patterns, and try to follow its conventions. What they cannot do is understand why those patterns exist. Code shows agents what shipped, not why one component, phrase, or interaction became your standard. That reasoning lives in design reviews, PR comments, Slack threads, and with the people who were in the room. For an agent, context that isn't in the codebase doesn't exist. Vercel is an agent-native team. We treat accepted product decisions like code, keeping them in the repository, reviewing changes against them, and making them available to every agent working there. The way we do this is through product-design. It's a system with three parts: An agent skill that gives coding agents the context behind decisions that require product or codebase judgment. Linters that enforce clear rules automatically. A review loop that gathers evidence from Slack, Figma, and GitHub, then prepares guideline updates for review. Any team can build the same structure around their own standards. Inside the product-design skill The skill lives inside the repository alongsid

AI SDK 7 is now available

June 25, 2026

AI SDK 7 is a major release for building production agents in TypeScript. The SDK has grown from model calls and chat primitives into a broader agent platform for developing, running, integrating, and observing agents across text, audio, realtime, image, and video. Every major provider is supported out of the box. At a Glance Develop agents with reasoning control, tool and runtime context, provider files and skills support, MCP Apps, and a terminal UI. Run agents with tool approvals, durable WorkflowAgent execution, first-class timeouts, and sandbox support. Integrate any agent harness such as Codex, Claude Code, Deep Agents, OpenCode, or Pi. Observe agent behavior with redesigned telemetry, @ai-sdk/otel, Node.js tracing channel support, lifecycle callbacks, and step performance statistics. Build beyond text with stable speech/transcription APIs, richer file parts, image generation and editing, multimodal embeddings, reranking, realtime voice (experimental), and video generation (experimental). Upgrade with clearer APIs through the new migration skill, codemods, and migration paths for ESM, Node.js 22, instructions, telemetry, stream, finalStep, runtimeContext, and toolsContext. Be

Deep Agents and OpenCode are now available in the AI SDK Harness

June 25, 2026

The AI SDK Harness lets you run established coding-agent runtimes through one unified interface, so you can switch runtimes without changing your application code. Today we're adding two new adapters, Deep Agents and OpenCode, both running inside a Vercel Sandbox. Deep Agents @ai-sdk/harness-deepagents adapts LangChain's deepagents runtime, with built-in file and shell tools, skills, host tools, multi-turn sessions, attach and resume, and built-in tool approvals. Read the Deep Agents harness documentation to get started. OpenCode @ai-sdk/harness-opencode boots a real OpenCode server inside the sandbox via @opencode-ai/sdk and streams its session events through the harness. It exposes OpenCode's built-in tools, supports both built-in and host tool approvals, and lets you pick the model, provider, and reasoning variant. Read the OpenCode harness documentation to get started. The full supported list of harnesses is now: Claude Code, Codex, Deep Agents, OpenCode, Pi, with more coming soon. Read more

Vercel Flags no longer requires SDK Keys for Vercel deployments

June 24, 2026

New projects using Vercel Flags no longer need to configure SDK Keys or the FLAGS environment variable when evaluating flags inside a Vercel deployment. At runtime, the Vercel adapter automatically receives a short-lived OIDC token, so authentication is handled for you with zero configuration. For local development, link your project with vercel link and pull credentials with vercel env pull. That's it. Existing projects and all SDK Keys are unaffected. This change only applies to new projects, and SDK Keys remain fully supported and are still required for: Cross-project flag access Non-Vercel runtimes Custom authentication setups Read the Vercel Flags documentation to get started. Read more

Pro teams can now run up to 500 concurrent builds

June 24, 2026

Pro teams can run up to 500 concurrent builds, up from 12. Teams with large repos can ship faster without waiting on queued builds. On-demand concurrent builds are enabled by default for Pro and Enterprise, and you're billed only for the build minutes you use. Learn more about on-demand concurrent builds. Read more

GLM 5.2 Fast via Wafer now available on AI Gateway

June 24, 2026

GLM 5.2 Fast via Wafer is now available on AI Gateway. Based on our own benchmarking across small-context, large-context, and tool-call scenarios, Wafer delivers a 2x higher throughput than other providers serving GLM-5.2 on serverless, leading on decode and end-to-end speed for sustained generation in the small- and large-context cases. In our testing, GLM 5.2 Fast on Wafer measured: Small context: 170+ tok/s Large context: 200+ tok/s To use GLM 5.2 Fast, set model to zai/glm-5.2-fast in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try GLM 5.2 Fast in the model playground. Read more

Deploy Node servers with zero configuration

June 23, 2026

You can now deploy a Node.js server to Vercel with zero configuration. Vercel detects a server.ts file at the project root or at src/server.ts and deploys it as a Node.js application, in addition to existing zero-configuration backends like Express, Koa, and NestJS: Vercel CLI can handle local development and deployment: Backends on Vercel are powered by Fluid compute with Active CPU pricing. Learn more about the Node.js runtime on Vercel. Read more

Vercel CLI now supports verifying DNS configuration

June 23, 2026

You can now use the Vercel CLI to diagnose a domain's DNS configuration. Using the vercel domains verify command, you can check whether a domain is correctly configured, attached to the right project, and verified. If not, it returns the exact steps to fix it. vercel domains verify inspects ownership, DNS records, nameservers, DNSSEC, conflicting records, and project attachment, and tells you precisely what to change. When the domain is correctly configured, you get a single line back: With --format json, the command returns the full diagnosis: status, reason, recommended records, detected conflicts, and suggested next commands. It exits non-zero when misconfigured or unverified, making it scriptable for CI and automation: You can also scope to a project, enforce exact matching, or run non-interactively: Upgrade your Vercel CLI to version 54.15.1 to get started. Learn more in the vercel domains verify documentation. Read more

Scaffold your chat apps with create-chat-sdk

June 23, 2026

Creating a new Chat SDK bot now takes a single command. create-chat-sdk scaffolds a complete Next.js project with your chosen platform adapters, a state adapter, environment variables, and a webhook route. The CLI walks you through selecting your platform adapters (e.g., Slack and Discord) and a state adapter (e.g., Redis), then installs the dependencies for you. Everything wired up: src/lib/bot.ts with onNewMention and onSubscribedMessage handlers, a populated .env.example, and a pre-configured next.config.ts. Pick your stack: combine any of the official or vendor-official platform and state adapters into a single project. The CLI installs only what you select. Fully scriptable: every prompt can be skipped with flags, so create-chat-sdk works with your AI agents and in CI. Coding agents are automatically detected, so the CLI runs non-interactively when they scaffold your application. For example, scaffold a Slack and Discord bot with Redis in one command: Read the documentation to get started. Read more

Preserve local environment variables when linking with the Vercel CLI

June 23, 2026

The Vercel CLI now preserves your .env.local file when running vercel link. Previously, linking could overwrite variables already in the file. The CLI now updates VERCEL_OIDC_TOKEN if it exists, or appends it if missing, without touching anything else. Run pnpm i -g vercel@latest to update, then run vercel link to get started. Learn more in the vercel link documentation. Read more

Custom OIDC Token Audiences

June 23, 2026

Vercel's OIDC issuer (oidc.vercel.com) now supports custom audiences. Deployments can request OIDC tokens with a specific audience claim, enabling secure service-to-service authentication with third-party providers. Why custom audiences? Vercel OIDC tokens are issued with a fixed audience (https://vercel.com/{owner}). While most cloud providers don't require a specific audience value, using a unique audience per provider is a security best practice. If a provider is compromised, an attacker cannot replay the token against a different provider - the mismatched aud claim will cause verification to fail. This new service makes it easy to mint provider-specific tokens without managing additional infrastructure. How it works When a Vercel deployment runs, it receives an OIDC token signed by Vercel. The new exchange service accepts this token and returns a new one signed with the same key, but with an updated audience (aud) claim targeting your downstream service. The exchanged token: Preserves all original claims (project, environment, owner, expiration) Sets the iss (issuer) to https://oidc.vercel.com/{owner}, scoped to the team that owns the deployment Includes an act (actor) claim wi

Deploy from Claude Design to Vercel

June 23, 2026

Vercel is now a send-to destination in Claude Design. When you finish a design, you can send it to Vercel and get a live URL back without leaving your canvas. Claude Design deploys the design as a new project in your connected Vercel account and returns a URL you can open and share. When you're ready to share a design, add Vercel as your destination in the “Share” menu and connect the Vercel MCP server to get started. Learn more about using Claude Design and Vercel together. Read more

Redesigned trace viewer for Vercel Workflows

June 23, 2026

The trace viewer for Vercel Workflows and Workflow SDK has been redesigned to better support inspecting runs from start to finish. Search across spans, zoom into any section of the timeline, and step through with the keyboard to find what you're looking for fast, then click into any step to see its inputs, outputs, and run metadata. The trace viewer is also available locally through Workflow SDK with npx workflow@beta web to inspect runs during development. Learn more about Vercel Workflows. Read more

Increased limit for projects per Git repo

June 22, 2026

Hobby users can now connect up to 25 projects per repository, up from 10. This makes it easier for users to onboard monorepos and use cases where one codebase maps to many deployable apps. Learn more about repository connection limits. Read more

WebSocket support is now in Public Beta

June 22, 2026

Vercel Functions can now serve WebSocket connections, enabling bidirectional communication between clients and server-side code on Vercel. Use WebSockets for realtime features such as interactive AI streaming, chat, and collaborative apps. WebSocket connections run on Fluid compute and follow the same limits and pricing as other Function invocations. With Active CPU pricing, billing only applies to the time your Function spends processing messages, not idle connection time. You can serve WebSocket connections using standard Node.js libraries, with no additional configuration: Higher-level libraries like Socket.IO are also supported. Read the documentation to get started. Read more

Sakana Fugu Ultra now available on AI Gateway

June 22, 2026

Sakana Fugu Ultra from Sakana AI is now available on AI Gateway. Fugu Ultra is built on a pool of publicly accessible frontier models, rather than running as a single model. It coordinates several models, routing work to 1-3 agents depending on the problem and combining their results into a single answer. Based on reasoning and scientific benchmarks, Fugu Ultra has capabilities similar to those of Claude Mythos Preview and Fable 5. To use Fugu Ultra, set model to sakana/fugu-ultra in AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Try Sakana Fugu Ultra in the model playground. Read more

Chat SDK adds Kapso support

June 22, 2026

Chat SDK now supports Kapso with the new vendor-official adapter. Kapso connects your bot to WhatsApp through its hosted platform, handling the WhatsApp Business setup, credentials, and webhooks so you can focus on your bot's logic. Replies use the standard Chat SDK thread and message APIs, with support for buttons and cards, media, reactions, contacts, and conversation history. The adapter maps each WhatsApp conversation to a Chat SDK thread, tied to a specific phone number and contact, and each inbound WhatsApp message to a Chat SDK message. Read the Kapso documentation to get started. Read more

Chat SDK adds Novu support

June 22, 2026

Chat SDK now supports Novu with the new vendor-official adapter. One handler set puts your agent on Slack, Microsoft Teams, WhatsApp, Telegram, and email. Novu handles credentials, identity, and delivery, keeping OAuth and tokens outside your app and mapping each channel to one user. Your agent always knows who they're talking to. Your agent can also send proactive notifications and handle the replies in the same loop, on whichever channel the customer used. One CLI command connects a real channel to your agent: It signs you in, creates a bridge agent, sets up the channel you pick (e.g., Slack), and adds the credentials to your project. Read the Novu documentation to get started. Read more

Chat SDK adds Sendblue support

June 22, 2026

Chat SDK now supports Sendblue with the new vendor-official adapter. Build bots that send and receive iMessage, SMS, and RCS through Sendblue's hosted gateway, reaching people on the messaging apps they already use. Messages use iMessage-first delivery with support for automatic SMS and RCS fallback, tapbacks, typing indicators, and delivery status callbacks. The adapter maps each Sendblue conversation to a Chat SDK thread, tied to a specific phone line and contact (or group), and each inbound iMessage, SMS, or RCS message to a Chat SDK message. Read the Sendblue documentation to get started. Read more

Chat SDK adds Linq support

June 22, 2026

Chat SDK now supports Linq with the new vendor-official adapter. Build bots that send and receive texts in both direct messages and group chats, with bidirectional media and native iMessage tapback reaction support. Replies use the standard Chat SDK thread and message APIs, with HMAC-verified webhooks and stable threading. The adapter maps each Linq chat to a Chat SDK thread, each text to a message, and each iMessage tapback to a reaction, so subscriptions, handlers, posts, and reactions work the same as with any other Chat SDK adapter. Read the Linq documentation to get started. Special thanks to Fardeem Munir, whose community contributions laid the groundwork for this adapter. Read more

Vercel Flags: Platform-native feature flags

June 22, 2026

At Vercel, feature flags are how we ship. From new features to model updates in v0, and even infrastructure changes like a production database migration where a flag was the cutover. The v0 team alone runs hundreds at any given moment. Merging code sends a build to production, but the feature flags control whether users can see what changed. Flags let you ship on your own schedule, release to segments when you're ready, and roll back immediately by toggling a flag, without touching source files or redeploying. Vercel Flags is platform-native: server-side by default, zero impact on page performance, and directly integrated with the frameworks you already use. What is Vercel Flags Vercel Flags lets you create feature flags, define targeting rules by user attributes, segments, or environment, run progressive rollouts, and flip kill switches if something breaks in production. From your code, you read flags through Flags SDK, an open-source, provider-agnostic library we maintain with first-class adapters for Next.js and SvelteKit. If you are using another framework, you can consume Vercel Flags using the built-in OpenFeature provider. The Vercel Flags dashboard sits alongside your proje

Workflow SDK now compresses run and step payloads

June 22, 2026

The Workflow SDK 5 beta now compresses all run, hook, and step inputs and outputs with zstd. Compression kicks in automatically, but only when it helps. Small payloads stay as-is, larger ones get compressed before they're persisted. Compressed payloads use less storage and are faster to read and write, so your workflows run faster and cost less. The savings are largest for JSON payloads typical of AI conversations, where storage size and cost can drop by up to 85%. One typical Workflow, run and stored without compression (52 MB): The same Workflow, run and stored with zstd compression (10 MB): Since eve builds durable agents on the Workflow SDK, the same compression now applies to the conversation history and state it persists for every session. That means eve agents store less and run faster, with no code to change. Update to workflow@5.0.0-beta.19 or later and learn more in the documentation. Read more

Vercel CLI now supports signing blob URLs

June 22, 2026

You can now generate signed URLs for Vercel Blob directly from the Vercel CLI. A signed URL is a scoped URL with a set expiration time that lets you perform a single operation on a specific object. Each URL is scoped to one operation (get, head, put, or delete), one pathname, and a custom expiry of up to 7 days. Update the Vercel CLI to version 5.14.5 to get started. Generate a presigned URL in one command Use the new vercel blob presign command to sign a URL for a single operation, for example to provide short-lived access to a private file or to allow your users to upload images of a certain size and content type. By default it issues a signed GET URL: Scope it to an upload with content-type and size constraints Control how long the URL stays valid with --valid-for (for example 15m, 1h, 7d) or an absolute --valid-until timestamp. Pass --json to get the result, operation, and expiry as structured output for scripting. Issue a standalone signed token Use vercel blob signed-token when you want to separate signed token generation from URL generation, for example to restrict write access to your private blob store while your agent investigates an issue. You can then feed the resulting

Introducing eve

June 21, 2026

Introducing eve — an open-source agent framework from Vercel for building, running, and scaling agents in production. Durable execution, sandboxed compute, approvals, channels, tracing, and evals built in. Like Next.js, for agents.

The Agent Stack

June 20, 2026

Agents are designed to do almost any kind of work, from answering support tickets to writing code. No matter how complex the workload, how long it runs, or how many turns it takes to complete, every agent needs three core capabilities to operate: Agents need to connect to models and route between them Agents need to run workflows across many steps Agents need to connect to the systems that make them useful and the platforms people use to interact with them Implementing these capabilities to build a complete agent forces developers to choose between vendor lock-in with a single provider API, stitching together solutions, or building abstractions themselves. The Agent Stack gives you all the building blocks you need to create and ship production-grade agents. Connect to models Agents don't run on a single model. Every task has a different cost, latency, and capability tradeoff, and the right call depends on what the agent is doing. It needs one interface to reach any of them, a way to route between them, and a way to stream back to the user. AI SDK gives an agent one interface to call any model, and AI Gateway routes across hundreds of them from a single endpoint. AI SDK Every lab ex

Vercel Connect: Secure access to external services for your agents

June 17, 2026

Vercel Connect lets your apps and agents access external services like Slack, GitHub, Salesforce, and your own custom APIs without storing a long-lived provider secret in your environment. You register a connector once, and your code requests scoped, short-lived tokens at runtime, only when it needs them. You can create a connector from your dashboard or the Vercel CLI: Your agent can then request a token from it at runtime, only when it needs it. The SDK fetches and refreshes that token automatically, so there is no secret for you to store or rotate by hand: Scoped tokens Each token is scoped to the task in front of it, with granularity that depends on the provider. A GitHub token can be limited to a single repository and read-only, instead of your whole organization. An agent gets exactly the access a task needs, never broad standing access to everything. A token can also be tied to a specific user instead of the app. It then acts with that user's identity and inherits their permissions, so the agent can do only what that user is allowed to do: Isolation and revocation A connector only works in the environments you attach it to, so you can run separate connectors for development,

Introducing eve, an open-source agent framework

June 17, 2026

eve is now available in public preview. eve is an open-source framework for building, running, and scaling agents. An agent is just a directory of files, and production comes built in: Durable execution Sandboxed compute Human-in-the-loop approvals Subagents Evals The smallest agent that runs is just two files, a model and a set of instructions. Add a tool, skill, channel, or schedule by adding a file. eve picks them up at build time and wires them in for you, so there's no boilerplate to register them. You can scaffold and start a new agent with a single command. It installs the dependencies, scaffolds the project, and starts a dev server, so you have an agent running locally in under a minute. Or you can let your coding agent set it up for you. Give it this prompt: And because an eve agent is an ordinary Vercel project, vercel deploy ships it to production unchanged, exactly as it ran on your machine. eve is the framework that Vercel builds and runs its own agents on. For the full tour, read the announcement or the documentation, and you can follow development in the open over at github.com/vercel/eve. Read more

Introducing Vercel Connect

June 17, 2026

Giving your agents access to your tools, data, and services is what makes them useful. As agents perform deeper work across systems, authenticating and authorizing that access becomes central to your application architecture. Today, agent access is usually granted through long-lived provider tokens stored in your environment variables, provisioned for everything your agent might need. These tokens are shared across every user, never expire, and give your agent full reach across every task, no matter how small the job. A vault makes that token harder to steal. It doesn't make it less dangerous. The problem is what happens when the token leaks: everything it can touch is now exposed. We built Vercel Connect to solve this problem. Now in Public Beta, Vercel Connect replaces the stored token with runtime credential exchange. You register a connector once. When your agent has work to do, your app proves its identity to Vercel Connect and gets back a short-lived credential, scoped to the task. Everything you used the token for still works. The agent just requests access each time instead of holding it. Register a connector once, then reuse it across projects and environments A connector

CLI deployment limits removed

June 17, 2026

We've removed CLI-specific deployment limits, making it easier to deploy from local machine and external CI/CD pipelines with instant feedback. Teams and AI agents can now deploy at the pace their workflows demand. Learn more about limits in the Documentation. Read more

Vercel Passport is now in Public Beta

June 17, 2026

Enterprise teams can now control access to their Vercel deployments with Vercel Passport, using their own identity provider. Visitors authenticate through providers like Okta, Auth0, or any compatible OIDC provider before they can view a protected deployment. Use Passport to: Reuse an OIDC application across multiple projects Set a team default that applies to new projects automatically Assign Passport to existing projects in bulk After Passport authenticates a visitor, Vercel injects a signed JWT into the x-vercel-oidc-passport-token request header. Read it server-side to access the external_sub claim, the stable visitor identifier returned by your identity provider: Read the documentation to get started. Read more

Vercel for Enterprise Apps and Agents

June 16, 2026

Today we are introducing Vercel for Enterprise Apps and Agents, a platform that gives your entire company the ability to ship with AI safely, behind your access and security boundaries. Over the past year, employees across Vercel shipped hundreds of agents and internal apps. Getting to production was the easy part, because we built them with eve on top of the Agent Stack and deployed them on Vercel. The difficult questions came after those agents were being used by our employees across the company: Who is allowed to use each agent? How do we keep internal agents internal? Which data and systems are agents allowed to touch? Which models are agents using, and how much are they costing us? Vercel for Enterprise Apps and Agents We built Enterprise Apps and Agents to answer these questions for ourselves. It makes ownership, access, and security the defaults your builders inherit instead of the projects your platform team queues. Platform components Security implementation Vercel Passport Puts every internal app and agent behind your identity provider by default Vercel Connect Gives agents short-lived, scoped credentials for the systems they use, like Slack, GitHub, Snowflake, Salesforce

Vercel Sandbox can now run for up to 24 hours

June 16, 2026

Vercel Sandboxes can run uninterrupted sessions for up to 24 hours (up from 5 hours). This new max duration unlocks workloads that require longer runtimes, such as large-scale data processing, E2E testing pipelines, and long-lived agentic workflows. Pair with persistent sandboxes to maintain durable state across extended runs. The 24 hour max duration is available on all Pro and Enterprise plans. Learn more about limits in the documentation and see how Vercel Sandbox duration and persistence work. Read more

GLM 5.2 now available on AI Gateway

June 16, 2026

GLM 5.2 is now available on AI Gateway. Built for long-horizon tasks, GLM 5.2 carries project-level engineering context across a single task, runs long-running tasks more reliably, and follows engineering standards more consistently. The context window for this model has been upgraded to 1M tokens, up from 200K in GLM 5.1. To use GLM 5.2, set model to zai/glm-5.2 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Workflow SDK now supports inflight cancellation

June 16, 2026

The Workflow SDK 5 beta now supports the standard AbortController and AbortSignal APIs across workflow and step boundaries. Create a controller inside a workflow, pass its signal into one or more steps, and cancel in-flight operations using the same API fetch already uses. That signal stays durable across suspensions and deterministic replay. When a step is running, it sees the cancellation, even when it's in a separate function invocation. Cancellation is also cooperative; steps have to inspect the signal or pass it to an API that supports AbortSignal. Use it to stop a slow step when a durable timeout wins a race, cancel the remaining requests after the first successful response, thread one signal through a multi-step pipeline, or cancel parallel work when an external condition changes. Try it with workflow@beta and read the cancellation documentation to learn more. Browser Workflow SDK resources in the Vercel Knowledge Base. Read more

Workflow SDK now supports TanStack Start

June 16, 2026

Workflow SDK now supports TanStack Start applications on Vercel. TanStack Start is built on Vite and Nitro, so the existing workflow/vite plugin works directly. Add it to vite.config.ts alongside tanstackStart(). From there, write workflow and step functions in standard TypeScript with "use workflow" and "use step". They run as durable, resumable operations that survive restarts, sleep for days, and retry on failure, with compilation, queue configuration, and persistence handled by the plugin. Read the TanStack Start guide to learn more and create your first durable workflow, or browse TanStack resources in the Vercel Knowledge Base. Read more

Vercel Functions can now run up to 30 minutes

June 15, 2026

Vercel Functions using the Node.js and Python runtimes now support execution durations up to 30 minutes for Pro and Enterprise teams, more than 2x the previous 800 second limit. Support for additional runtimes is coming soon. Use longer-running Functions for work that needs more time to finish, including: Long LLM reasoning and tool calls AI responses that stream for several minutes Document and media processing OCR and extraction Web scraping and browser automation Complex Workflow steps or Queue handlers Fluid compute keeps long-running work cost-efficient. Active CPU billing only applies while your code is executing, and pauses while your Function is waiting on I/O such as AI model calls, database queries, and third-party APIs. Set maxDuration to opt in. For Next.js App Router, configure it in the route file: For other runtimes and frameworks, configure maxDuration for a specific function path in vercel.json: Durations above 800 seconds are in beta and require Fluid compute. Learn more about configuring max duration for Vercel Functions in the documentation. Read more

Auth0 joins the Vercel Marketplace

June 15, 2026

You can now add Auth0, a production-ready authentication to your Vercel app in just a few clicks. Built for modern frameworks like Next.js, Auth0 is an identity and access management platform for securing your apps and agentic workflows. This integration enables: Automatic provisioning of an Auth0 application that connects to your Vercel project Out-of-the-box support for your Next.js applications using the Auth0 Next.js SDK Complete user management with hosted dashboards, sessions, and roles Sync authentication configuration across Development, Preview, and Production environments Get started with Auth0 on the Vercel Marketplace. Read more

Chat SDK now supports rich text in Telegram

June 15, 2026

Chat SDK now renders explicit markdown and ast messages as native rich messages on the Telegram adapter. Your bots get real headings, lists, tables, task lists, formulas, and separate media blocks instead of flattened text. What you get: Native formatting: headings, lists, tables, task lists, formulas, and media blocks. Streaming: private chats show live draft previews and persist the final response. Automatic fallback: plain strings, raw messages, cards, and captions continue to work exactly as before. Read the Telegram adapter documentation to get started. Read more

Increased Blob store limit for Hobby users

June 15, 2026

Hobby users can now create up to 100 Blob stores, up from 5. This gives teams more flexibility to organize data by project, environment, or region as applications grow. Storage, operations, and transfer limits still apply. Learn more in the Blob documentation. Read more

Workflow SDK now runs natively in Nitro v3

June 13, 2026

Workflow SDK's native Nitro v3 integration is now in beta. Steps run inside the same bundled runtime as the rest of your app, instead of a separate bundle. Nitro's useStorage() and other server-side APIs work directly inside "use step" functions. The Nitro dev server also serves the workflow web UI at /_workflow. Open it in your browser to inspect, monitor, and debug workflow runs. Workflow routes are now bundled by Nitro as part of the app build. Dependencies are traced, and unused code is tree-shaken, so the output includes only what runs, with faster builds and smaller bundles. Get started with Workflow SDK on Nitro. Read more

Program Claude Code, Codex, Pi and other agent harnesses with AI SDK

June 12, 2026

AI SDK 7 introduces HarnessAgent, a single API for running established agent harnesses, including Claude Code, Codex, and Pi. AI SDK has always let you switch models without rewriting your agent. Now you can switch the harness the same way. Write the agent once. Use the best harness available. Today. In 3 months. A year from now. Harnesses manage the components above a model call, including skills, sandboxes, sessions, permission flows, compaction, runtime configuration, and sub-agents. The AI SDK normalizes access to those capabilities through a unified harness abstraction. Initial harness adapters for this experimental release include Claude Code, Codex, and Pi, with more coming soon. Swap claudeCode for codex or pi and keep the same HarnessAgent flow. Every harness runs the agent in a sandboxed workspace, keeping the host environment safe. Both HarnessAgent.generate() and HarnessAgent.stream() return AI SDK-compatible results. If your app already uses useChat or related AI SDK tooling, you can swap in HarnessAgent without changing your user interface code. HarnessAgent is available on the AI SDK canary release. Read the AI SDK harness documentation to get started. Read more

Introducing Vercel Drop

June 12, 2026

Vercel Drop lets you deploy a file or folder by dragging it into your browser. You don't need Git, the Vercel CLI, or any local setup. Drop a project onto vercel.com/drop, pick a team and project name, and select Deploy. Vercel will create a new project, upload your files, and publish them straight to production with a live URL you can share. All in a matter of seconds. Vercel Drop handles more than static files: - Framework projects: Vercel detects your framework (e.g., Next.js) and builds it. Exports from tools like Bolt.new deploy this way. - Static sites: Files with no framework deploy as-is, with no build step. That includes exports from Claude Design and Google Stitch. If your folder has no index.html at the top, you choose which page loads at your site's root. Each drop creates a new project. To get automatic deployments on every push, connect a Git repository to the project afterwards. Get started at vercel.com/drop or read the documentation. Read more

Claude Fable 5 access suspended on AI Gateway

June 12, 2026

In compliance with the US Government's directive, access to Claude Fable 5 is now suspended on AI Gateway for all users. We do not know if or when access to the model will be restored. You can still access and use other Anthropic models through AI Gateway. Read more

Kimi K2.7 Code now available on AI Gateway

June 12, 2026

Kimi K2.7 Code from Moonshot AI is now available on AI Gateway. K2.7 Code is a coding model built for long-horizon programming tasks, generalizing across scenarios including frontend development, DevOps, and performance optimization. The model has a native multimodal architecture that supports text and vision input, and always runs in thinking mode. To use K2.7 Code, set model to moonshotai/kimi-k2.7-code in the AI SDK: Pass an image alongside a prompt to use the model's multimodal input: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

How Okara runs CMO agents for 120,000 companies on Vercel

June 11, 2026

Okara on Vercel 4 billion tokens processed daily across a multi-provider AI stack on Vercel AI CMOs actively managing growth for 120,000+ businesses Eight sub-agents handling SEO, GEO, social, content, Reddit, and Hacker News New AI models available to users the same day they ship Okara is an AI CMO that directs a team of specialized sub-agents to drive marketing, so founders don't have to. Give Okara your website URL, and the AI CMO builds a marketing strategy, develops a brand voice, and activates agents across SEO, content, and social media to drive awareness and pipeline without a single marketing hire. As Fatima Rizwan, Okara's founder, puts it: "You can build something in a weekend and spend months trying to get anyone to notice." Distribution, she argues, is stuck in the pre-AI era: fragmented across subscriptions and agencies that cost over $15,000 a month before a single dollar comes back. Okara is built by a team of four and processes 4 billion tokens a day. The company operates on the new model for startups: a small team building a platform that handles growth for thousands of other companies. Fatima quickly learned that serving hundreds of thousands of companies with fo

Vercel plugin is now available in Grok Build

June 11, 2026

The Vercel plugin is now available in Grok Build. Grok can now draw on Vercel knowledge as you work. Real-time activity, including file edits and terminal commands, dynamically injects the relevant knowledge into context, so answers stay aligned with current platform APIs and recommended patterns. Install it in either of two ways: Add vercel to your prompt and Grok will recommend installing it in chat Open the Grok Build marketplace with /marketplace, then search for and install Vercel Learn more about the Vercel plugin in the documentation. Read more

DeepSeek models now available via Azure on AI Gateway

June 11, 2026

Azure is now a provider for DeepSeek V4 Pro and V4 Flash on AI Gateway. Requests to either model can route through Azure alongside the existing providers for another failover path. No code changes are required: default routing considers Azure automatically, and if a provider fails the gateway falls back through the remaining list. If you want requests to try Azure first, use order in the gateway provider options to prefer Azure while keeping the other providers as fallback for deepseek/deepseek-v4-pro or deepseek/deepseek-v4-flash in the AI SDK: If you have existing Azure credentials, you can bring your own key and AI Gateway will use it for requests routed to Azure. See API key authentication and BYOK for setup. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, budgets for API keys, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway m

Threshold billing is now enabled for Pro teams

June 10, 2026

Threshold billing now sends Pro teams a partial invoice mid-cycle once on-demand usage reaches a threshold, instead of holding all charges until the end of the billing period. Partial invoices and the end-of-cycle invoice add up to your total usage, so the same usage is never billed twice. Learn more about partial invoices. Read more

Claude Fable 5 now available on AI Gateway

June 9, 2026

Claude Fable 5 from Anthropic is now available on AI Gateway. A Mythos-class model, Fable 5 is a notable step up over prior Claude models on long-running, ambiguous, multi-step tasks, executing end-to-end on work that previously required frequent human check-ins. The model sustains productive output across multi-day runs and dependably dispatches parallel sub-agents, and lower effort settings often match what prior Claude models produced at their highest effort. Code review, bug-finding, and repository investigation are stronger, and first-shot correctness on complex problems is noticeably higher. Fable 5 ships with blocking classifiers in place that refuse offensive cybersecurity, biology, and summarized-thinking extraction, because the model's capabilities in those areas introduce real misuse risk. Anthropic also does not support Zero Data Retention because some misuse patterns only become visible across cumulative requests. Prompts and completions are retained for 30 days and are not used to train Claude. To use Fable 5, set model to anthropic/claude-fable-5 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, fai

How Code and Theory cut time-to-prototype 75% with v0

June 9, 2026

Code and Theory on Vercel Time-to-prototype cut by 75% Deployment timelines reduced by over 50% Wireframe and PRD process replaced by live prototypes Code and Theory is a digital-first creative and technology agency that combines strategy, design, and engineering for brands like Microsoft, Amazon, and NBC. The agency replaced wireframes and requirement documents with prompt-to-code workflows in v0, so product, engineering, and go-to-market teams can turn a client brief into a working prototype in real time. Time-to-prototype dropped by 75%. Read more

How Fern runs multi-tenant docs for Webflow and ElevenLabs on Vercel

June 9, 2026

Fern on Vercel 3x faster time to first byte Page load times reduced by 80% 6 million+ page views per month from 1 million+ unique visitors 65% of the platform migrated from Pages Router to App Router in 7 days Fern helps companies ship developer documentation and SDKs, running customer docs for Webflow, ElevenLabs, and others across custom domains from a single Next.js app on Vercel. Mid-migration from the Pages Router to the App Router, the team moved 65% of the platform in seven days instead of taking on a months-long engineering project. Today Fern deploys multiple times a day, with deploys down to five minutes and page load times cut by up to 80%. Read more

Budgets for API keys on AI Gateway

June 9, 2026

AI costs are getting harder to forecast. As teams lean more on coding agents and other token-heavy workflows, a key can burn cost faster than anyone notices: Autonomous workflows that can loop or fan out without supervision Demos and prototypes that could catch unexpected traffic if shared or shipped Developers exploring or experimenting without a sense of per-model cost Set a spend cap on any key, and AI Gateway rejects further requests on that key once the limit is exceeded, until the budget resets or you raise it. The cap applies to all AI Gateway providers and models running through the key, making it easier to consolidate and govern AI costs. API key budgets in the Vercel Dashboard On the AI Gateway API Keys page, click Create Key, enable the Spend Quota option, enter a limit in dollars, and choose a refresh period. You can also edit existing keys and add, change, or remove budgets by clicking the right hand side ... menu and Edit Key. API key budgets in the Vercel CLI Create a budgeted API key programmatically via the Vercel CLI. The format is: Pair a key with an optional refresh period (daily, weekly, monthly, or none) to scope the limit to a window. Each period resets at th

Domain Search is now available through the Vercel CLI

June 9, 2026

You can now use the Vercel CLI to search domains. Using the vercel domains search command, you can supply a domain name and retrieve availability and price results for all TLDs that Vercel supports. You can also filter by TLD, apply sorting, and filter out unavailable domains. Upgrade your Vercel CLI to version 54.10.1 to get started. Read more

How the Weather Company serves real-time forecasts to 350 million daily active users on Vercel

June 9, 2026

The Weather Company on Vercel 350 million monthly active users served worldwide 2.2 billion coordinates calculated every 15 minutes 80% increase in velocity from design to live web page The Weather Company delivers weather forecasting and content to over 350 million monthly active users worldwide, backed by about 130 meteorologists tracking weather systems across the globe. By rebuilding its web serving stack and CMS on Vercel and v0, the team turned work that once took years into months and shortened the path from a design to a published website from days to hours. Read more

DeepSeek enters the fight for token volume, Anthropic continues to dominate spend

June 8, 2026

AI Gateway Production Index — June 2026 Every month, AI Gateway routes tens of trillions of tokens between production applications and AI labs, giving us visibility into what AI usage actually looks like, separate from leaderboards and benchmarks. We publish the data monthly in the AI Gateway production index. June 2026 summary The July index reports on AI Gateway data collected in May 2026. Total AI Gateway tokens grew +20% MoM; total spend grew +43% MoM. Customers paid almost 20% more per token on average than in April. DeepSeek’s share of tokens jumped from under 1% to 17% in a single month, while its share of spend stayed near 1%. Anthropic’s share of spend grew from 61% to 65% in May, holding 70–80% of spend across every high-stakes use case (AI app generation, back office agents, and coding agents). Cost-consciousness meant smarter routing between low-cost and frontier models. Customers got more deliberate about which model did which work, while overall usage kept climbing. Last month, headlines about blown token budgets dominated tech news: Uber burned through its annual Claude Code budget shortly after Q1 and Amazon shut down KiroRank to curb unproductive tokenmaxxing. Whil

The skills.sh API is now available

June 5, 2026

The skills.sh API is now available. Authenticate with your project's Vercel OIDC token and start querying more than 600,000 skills from across the open-source ecosystem. Search for skills, pull detailed info on any one, check its security audit, and more. Vercel issues a short-lived token scoped to your team and project, rotated automatically, so there's no long-lived secret to leak or rotate. On each request, skills.sh verifies the token and applies a rate limit of 600 requests per minute per team and project. Read the skills.sh API documentation to get started. Read more

Drives for Vercel Sandbox in Private Beta

June 5, 2026

Vercel Sandbox now supports drives in private beta. Drives are persistent, attachable storage with a lifecycle independent from any sandbox. Create a drive once, then mount it at a configurable path when starting a sandbox. When the sandbox stops, the drive remains available to attach to a later sandbox. Install the beta SDK (@vercel/sandbox@beta) or beta CLI (sandbox@beta), then create and mount a drive: Sandbox Drives are useful for: Keeping agent workspaces across disposable sandboxes Retaining cloned repositories, dependencies, and build outputs Managing data independently from the sandbox lifecycle During the private beta, a drive can be mounted read-write by one sandbox at a time. Sandbox drives should not be used for production data while in private beta. Sign up here to join the waitlist, learn more in the docs, or read the complete guide. Read more

Updates to Legal Terms

June 4, 2026

The proliferation of agentic workflows means developers now regularly grant AI tools direct access to their infrastructure, use services that act autonomously, and build on platforms that themselves use AI to operate. We’ve updated our Terms of Service and Marketplace terms to clarify shared responsibility when actions on your account may be taken by AI, whether Vercel's own or a third-party tool you've connected, as well as other important updates detailed below. What is changing? AI Functionality and Third-Party Tools Vercel's platform increasingly includes AI-powered services, including v0, Vercel Agent, WAF natural language rules, and AI Gateway. These services may take actions on your behalf. We've introduced two new concepts to make the responsibilities around this explicit: AI Functionality refers to Vercel services that are AI-powered, that help you build AI products (like AI SDK), or that connect to AI providers (like AI Gateway). Third-Party Tools refers to any non-Vercel tool, script, service, or automation (AI-powered or not) that you grant access to your Vercel account. This includes CI/CD pipelines, third-party agents, and integration scripts. The key responsibilities

Nemotron 3 Ultra now available on AI Gateway

June 4, 2026

Nemotron 3 Ultra from Nvidia is now available on Vercel AI Gateway. Nemotron 3 Ultra is an open Mixture-of-Experts reasoning model built for orchestrating long-running agent workflows, with a 1M token context window. The model targets multi-turn agent workflows: planning, tool use, sub-agent delegation, and error recovery. Throughput reaches up to 350 tokens per second, with up to 30% lower cost on agentic tasks. To use Nemotron 3 Ultra, set model to nvidia/nemotron-3-ultra-550b-a55b in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, dynamic provider sorting by latency and cost, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Build and deploy Shopify storefronts on Vercel

June 4, 2026

You can now create a Shopify store directly from Vercel and start building a production-ready storefront in minutes. Install the Shopify integration to automatically configure your Shopify credentials in Vercel. Create a free test store, build with v0 and deploy without leaving your workflow. When you're ready to launch, you can claim the store and take ownership of it. Update: Connect an existing Shopify store to Vercel is now supported. Get started by installing Shopify from the Vercel Marketplace, v0 , or Vercel CLI to start building your next commerce experience. Read more

Trace any Vercel request from the CLI

June 3, 2026

You can now generate Session Traces through the Vercel CLI. Use the new vercel curl --trace command to generate an OpenTelemetry trace to the specified endpoint from the terminal. Use the new vercel traces get command to fetch the generated trace by request ID. Available on all plans. Update the Vercel CLI to the latest version and run vercel curl --trace to get started. Learn more about the vercel curl CLI command. Read more

Grok Imagine Video 1.5 on AI Gateway

June 3, 2026

Grok Imagine Video 1.5 from xAI is now available on AI Gateway. The model generates video from an input image with synchronized audio in a single pass. This release improves audio quality, prompt following, and photorealism. Face accuracy and character consistency are stronger across longer sequences, with better lighting and physical realism in the output. Reference image support has been expanded to give more control over visual style and subject. To use this model, set model to xai/grok-imagine-video-1.5-preview in the AI SDK. Chain an image model with Grok Imagine Video 1.5 to generate a still and animate it in one flow: You can also try this Grok Imagine Video 1.5 directly in the AI Gateway Playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, dynamic provider sorting by latency and cost, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway and

Chat SDK adds Twilio adapter support

June 2, 2026

Chat SDK now supports Twilio, extending its single-codebase approach to Slack, Discord, GitHub, Teams, Telegram, and WhatsApp with the new Twilio adapter. Teams can build SMS and MMS bots that handle direct messages, inbound media, and message history. The adapter also ships low-level voice helpers, so you can build an app that handles voice calls using TwiML primitives. Modals, buttons, and reactions are not supported, and cards display as plain text. Outbound media must link to a URL accessible by Twilio. Read the documentation to get started, browse the directory, or build your own. Read more

Build Chat SDK web UIs in Vue or Svelte

June 2, 2026

The Chat SDK web adapter now has first-class support for Vue and Svelte, joining the existing React integration. Because the adapter speaks the AI SDK UI message stream protocol, the same server route works. Each framework ships its own useChat, built on the matching AI SDK package for React, Vue, or Svelte. The Vue and Svelte helpers return a reactive Chat instance, so read chat.messages and chat.status directly in your template rather than destructuring. The api and threadId options match the React wrapper, and your server setup remains the same. Read the documentation to get started or browse the adapters directory. Read more

Build custom Slack runtimes

June 2, 2026

Chat SDK now ships the Slack adapter's primitives as standalone imports for apps that already handle their own routing, state, or workflow execution. Use only what you need: Request verification and payload parsing (@chat-adapter/slack/webhook) Markdown formatting (@chat-adapter/slack/format) Fetch-based Web API calls (@chat-adapter/slack/api) Block Kit conversion (@chat-adapter/slack/blocks) Each subpath skips the full Chat runtime, so your imports stay clean. To get started, read the Slack primitives documentation. Read more

Signed URLs are now available for Vercel Blob

June 2, 2026

You can now generate time-bound signed URLs for Vercel Blob. A signed URL is a scoped URL with an expiry that allows you to upload, download, inspect, or delete a specific object without giving access to your entire Blob store. Each URL is scoped to a single operation (put, get, head, or delete), a single pathname, and an expiry you choose, up to 7 days. The signature covers the operation and constraints, so a URL signed for a GET can't be reused as a PUT. Direct uploads from the browser Upload URLs (put) support multipart, so the browser can stream large files straight to Blob storage without round-tripping through your server. Conditional deletes Delete URLs accept an ifMatch option so the delete only applies if the object hasn't been overwritten since you signed the URL: Signed URLs work alongside OIDC. Your server authenticates to Blob via OIDC, generates a signed token, and produces narrowly scoped, time-bound URLs for the browser, so your long-lived BLOB_READ_WRITE_TOKEN never leaves the server. Update @vercel/blob to 2.4.0 and read the documentation to get started. Read more

Edit Git settings for all projects in a repo

June 2, 2026

Monorepos that deploy many projects can now configure all of their project's Git settings more conveniently. Previously, if you wanted to consistently configure each project's settings for commit status, repository_dispatch events, etc., you had to click through to every project's settings and consistently apply the same setting. Now, you can do it all in one place. Try it out in project settings or visit the docs to learn more! Read more

Chat SDK adds Velt support

June 1, 2026

Chat SDK now supports Velt with the new vendor-official adapter. Build bots that read and reply within Velt comment threads, right where your team already works: documents, text editors, and canvases. Tag the bot, and it will answer in the same thread, grounding its reply with the document name, URL, and anchored text. The adapter maps Velt documents to Chat SDK channels, comment annotations to threads, and individual comments to messages. Read the Velt documentation to get started. Read more

Chat SDK adds AgentPhone support

June 1, 2026

Chat SDK now supports AgentPhone with the new vendor-official adapter. Give your bot its own phone number so it can handle voice calls and text messages using the same handlers you already write. When a call ends, the transcript is delivered as a message, allowing your bot to respond via SMS or iMessage. Calls and messages to the same number arrive via a single webhook, so you handle both from a single endpoint rather than a separate pipeline for each channel. They also land in one thread, so your bot maintains context throughout the conversation. Read the AgentPhone documentation to get started. Read more

Qwen 3.7 Plus now available on AI Gateway

June 1, 2026

Qwen 3.7 Plus from Alibaba is now available on Vercel AI Gateway. Both Qwen 3.7 Plus and 3.7 Max are free for paid AI Gateway users till 6/4/26 12:00pm PT. The model unifies vision and language into a single agent foundation, with capabilities spanning GUI and CLI operation, coding and productivity workflows with full-modality input, and visual agent tasks including perception and reasoning. It is designed to generalize across diverse agent harnesses. To use Qwen 3.7 Plus, set model to alibaba/qwen-3.7-plus in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, dynamic provider sorting by latency and cost, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Vercel Blob now supports OIDC authentication

June 1, 2026

Vercel Blob now supports OIDC authentication and is the default setting when connecting new projects. Vercel-issued OIDC tokens are short-lived and rotate automatically, so you no longer need a long-lived BLOB_READ_WRITE_TOKEN. To upgrade an existing store, first update your project to use the latest @vercel/blob, then navigate to the Projects tab under your Blob store and select Upgrade to OIDC from the project's context menu. Functions running on Vercel receive the token automatically and authenticate requests with it: The Vercel CLI picks up the same environment variables once you update it, so you and your agents can read and write to a private store from your terminal without a long-lived token: Read the documentation to get started. Read more

Elastic Build Machines now protect against out of memory builds

June 1, 2026

Elastic build machines now monitor your build's memory usage and automatically adjust to prevent out-of-memory (OOM) failures: If your build is fast but memory-intensive, we will no longer downgrade you to a smaller machine If your build is close to running out of memory, we will automatically upgrade to a higher tier If your build fails due to an out-of-memory error, the next deployment will automatically run on a higher tier Thresholds are set conservatively to balance deployment reliability and cost. Vercel only considers your build's memory usage, not the memory used by Vercel's own build infrastructure. Enable elastic builds in your team settings or project settings, or read the docs. Read more

Chat SDK adds Lark and Feishu support

May 31, 2026

Chat SDK now supports Lark and Feishu via a new vendor-official adapter. Build bots that post, edit, and delete messages, stream replies via Lark's native cardkit typewriter API, send interactive cards, and react with emojis across both Lark and Feishu conversations. The adapter connects over Lark's WebSocket transport, so bots run from any long-running Node process without exposing an HTTP webhook endpoint. To get started, read the Lark or Feishu documentation. Read more

MiniMax M3 on AI Gateway

May 31, 2026

MiniMax M3 is now available on Vercel AI Gateway. M3 is MiniMax's first model with a 1M-token context window and native multimodality, built around MiniMax Sparse Attention (MSA). M3 improves on software engineering, terminal-based tool use, and agentic web browsing, and is tuned for multi-turn collaboration. To use MiniMax M3, set model to minimax/minimax-m3 in the AI SDK. Pass an image alongside a prompt to use M3's multimodal input: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, dynamic provider sorting by latency & cost, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Function invocations now billed per unit

May 29, 2026

Function invocations are moving from package-based to per-unit pricing for Pro and new Enterprise customers. You’ll continue paying the same effective rate until the end of your current billing cycle. Starting with your next billing cycle you’ll be billed per unit to align costs directly with your usage. The new rate is $0.0000006 per invocation (previously $0.60 per 1M invocations) for Pro customers. Per‑unit billing scales more smoothly across team sizes and usage patterns. It also helps teams on Pro use function invocations without immediately consuming a large portion of the included monthly usage credit. Get started or learn more about Vercel Functions usage and pricing. Read more

Protecting against token theft

May 29, 2026

HTTP requests are inexpensive. Vercel charges ~$2/million, a fraction of a cent per call. But a single prompt to an agent on a frontier model can cost $2, making AI a million times more expensive, and inference theft one of the highest-margin businesses an attacker can run. We have seen this type of attack on our own APIs. If you have AI endpoints exposed to the internet, the risk of abuse is high and can easily run up bills in the tens of thousands of dollars or more. Protecting those endpoints requires verification to run on every AI request, not on the session or signup. Rate limits and auth walls aren't sufficient on their own because checks that run once per session get amortized away across thousands of stolen calls. At Vercel, we gate every AI request through BotID deep analysis, and you can do the same on your own endpoints with a few lines of code. What inference theft is Inference theft is the unauthorized use of someone else's paid AI inference, either for free consumption or downstream resale. The operator pays per AI call; the attacker pays nothing for inference and then resells the tokens at a discount. This goes beyond rate-limit abuse to actual resale of a stolen re

Run Docker containers inside Vercel Sandbox

May 29, 2026

Update (August 2026): Vercel Sandbox now runs on Vercel Managed Images. Starting with SDK v3, new sandboxes default to the Ubuntu-based vercel/sandbox/universal:latest, where dnf isn't available. The example below applies to the deprecated Amazon Linux runtimes; on Ubuntu, install Docker with apt-get. See the announcement for details. Vercel Sandbox now supports installing and running Docker inside a sandbox. An agent can build containers, install system packages, and modify files without touching your host system. Install Docker, start the daemon, and serve a containerized application: Docker in a Sandbox is useful for running containerized services like Redis or Postgres as test dependencies, validating container images before deploying, or previewing applications served from a container. Combined with persistent sandboxes, the Docker installation and pulled images carry over between sessions. As well as adding support for Docker, sandboxes now support FUSE filesystem drivers and VPN clients, unlocking unlimited capabilities to what can be built. Learn more about these new system specifications in the documentation. Read more

Port 8080 is now available in Vercel Sandboxes

May 29, 2026

Vercel Sandboxes now allow opening and binding to port 8080 as an ingress domain. This port was used as a controller port, and that has now moved to port 23456. Learn more about Sandbox in the documentation. Read more

Amazon OpenSearch Serverless is now available in the Vercel Marketplace

May 28, 2026

Amazon OpenSearch Serverless is now available in the Vercel Marketplace, with guided setup and automatic project configuration built into the Vercel dashboard. Built for agentic infrastructure By bringing Amazon OpenSearch Serverless into the Vercel Marketplace, teams benefit from a unified workflow: In-dashboard provisioning: Spin up OpenSearch collections directly from the Vercel dashboard Automatic configuration: Environment variables are automatically injected into your projects on provisioning Unified management: View and manage search resources alongside the rest of your application infrastructure in the Marketplace resource view, with quick links to AWS for deeper configuration This integration is powered by the next generation of Amazon OpenSearch Serverless APIs, with Vercel participating as a key design partner alongside AWS during development. Unified support for vector, lexical, hybrid, and agentic search in a single collection Autoscales up to 20× faster, built for the bursty, unpredictable load patterns of agentic workloads Scale to zero with no idle costs Up to 60% cost savings by paying for actual usage instead of peak capacity Get started To help you get started, y

Opus 4.8 on AI Gateway

May 28, 2026

Claude Opus 4.8 is now available on Vercel AI Gateway. Claude Opus 4.8 is built for long-horizon agentic execution and handles complex, multi-step coding tasks like refactors that previously required human correction mid-task. The model also produces clearer, less hedgy prose for knowledge work like drafting documents, analyzing data, and building presentations. To use Opus 4.8, set model to anthropic/claude-opus-4.8 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, Zero Data Retention support, dynamic provider sorting by latency & cost, and more. AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including on Bring Your Own Key (BYOK) requests. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Team-wide provider allowlist on AI Gateway

May 28, 2026

AI Gateway now supports a team-wide provider allowlist. Teams can restrict which providers can serve requests, so traffic only routes to approved providers. The allowlist applies to every request through AI Gateway, including Bring Your Own Key (BYOK) traffic. Regulated teams typically vet AI providers across multiple dimensions with security and legal sign-off, ending up with a vendor set that reflects the specific requirements of their org. The allowlist turns that approved-vendor list into a routing guarantee: Enforcement happens at the gateway level, not at the request level. A developer on the team cannot route traffic to a provider the org hasn't approved. This restriction also applies to coding agents. Even if an agent omits or modifies request-level provider filters, AI Gateway still blocks unapproved providers. Only team owners can modify the provider allowlist, keeping control centralized and auditable. New providers are disabled by default once the allowlist is on, so the approved set doesn't silently expand when AI Gateway integrates a new vendor. How to configure Toggle on Provider Allowlist in the AI Gateway Settings tab. All current providers are allowed by default,

How Conductor moved parallel coding agents from the laptop to the cloud with Vercel Sandbox

May 27, 2026

Conductor on Vercel Multiple parallel coding agents running at scale in the cloud Model-agnostic, works with Claude Code, Codex, and other coding agents Used by engineering teams at Notion, Linear, Ramp, and Life360 Interacting with a single coding agent feels natural, and Conductor makes directing a fleet of agents just as intuitive. It's a GUI for spinning up multiple coding agents in parallel, each working on an isolated branch of your codebase at once. You review their work, merge what works, and redirect the agents that need to keep iterating. Engineering teams loved it, but hardware was a real limitiation. Every agent ran on your machine, so the fan climbed, the CPU ran hot, and if you closed your laptop, the agents stopped. Charlie Holtz, co-founder and CEO of Conductor, spent time at Vercel before starting the company. He understands what developers lose when their environment imposes rules on how they can work. Cloud Workspaces, Conductor's remote execution layer built on Vercel Sandboxes, removes the local hardware constraint. Developers can now spawn multiple agents and close the lid without interrupting work. The journey from local to cloud Developers who adopted Conduc

Experimental native binaries for Vercel CLI

May 27, 2026

The Vercel CLI now ships an optional experimental native binary that starts faster, is even more secure, and requires no Node.js runtime dependency. Binaries are code-signed, allowing your OS to verify that they came from Vercel and haven't been modified. Additionally, on macOS, credentials are stored in the system Keychain scoped to the binary, so other processes cannot access them without explicit permission. You can opt in by installing the experimental package: The -f flag is required because @vercel/vc-native installs the same global vercel and vc bin names as the standard CLI. Once installed, vercel and vc run the native binary across macOS, Linux, and Windows on x64 and arm64. Let us know what you think on GitHub. Read more

Redesigned Deployments List

May 27, 2026

The deployments list has been redesigned with a denser layout, so you can see more deployments at once. Environments are now grouped with statuses, and the updated layout makes branches and commits easier to scan. The mobile experience has also been improved, making it easier to scan deployment activity on busy projects. Open your dashboard to see the updated design. Read more

Microfrontends routing now applies to vc alias and branch domains

May 26, 2026

This week we are gradually rolling out an update to Vercel Microfrontends routing for aliases and branch-assigned domains. Aliases inherit Microfrontends routing Aliasing a Microfrontends URL with vc alias now preserves the full microfrontends routing config from the source deployment. Previously, the new alias only inherited the deploymentId. Update to the latest Vercel CLI to pick up the change. Branch domains route across all projects A project domain assigned to a git branch now routes to that branch in every project in the Microfrontend that shares the branch name. Previously, the domain only routed to that branch within the project that owned the domain. See the Microfrontends documentation for details on routing, aliases, and domain assignment. Read more

Sandbox persistence is now GA

May 26, 2026

Vercel Sandboxes now automatically save and restore filesystem state between sessions. Persistence is on by default, meaning no snapshots to manage or state to track manually. Each sandbox has a durable, customizable name that acts as a unique reference in your project. You can create, retrieve, or resume a sandbox by name. Vercel spins sessions up and down automatically, without interrupting your workflow. Create a persistent sandbox When you call Sandbox.create(), persistence is enabled by default: Each automatic snapshot consumes snapshot storage, which is billed separately from compute. For ephemeral workloads, opt out of persistence to minimize storage costs: To opt out of persistence with the CLI, pass --non-persistent to sandbox create. Non-persistent sandboxes discard their filesystem when the session ends. Resume a persistent sandbox Resuming is automatic. Any call on a stopped sandbox, like runCommand() or writeFiles(), starts a new session from the most recent snapshot. Other improvements Sandbox.fork(): Create a new sandbox from an existing one Sandbox.getOrCreate(): Idempotent retrieve-or-create for long-lived sandboxes Sandbox.delete(): Permanently delete a sandbox Ri

Vercel Domains now supports price sorting and availability filtering

May 26, 2026

Vercel Domains now supports price sorting and availability filtering. Price sorting shows lower cost domains first, so you can quickly find a domain that fits your budget. Availability filtering moves unavailable domains to the bottom of the search results, so you can focus on domains that you can actually purchase. Try it at vercel.com/domains. Read more

Firecrawl joins the Vercel Marketplace

May 26, 2026

Firecrawl is now available on the Vercel Marketplace, allowing Vercel teams to power AI agents and applications with structured web data without managing crawling infrastructure. This integration helps developers scrape websites into LLM-ready formats, search and retrieve full page content, and interact with dynamic pages for retrieval and agent workflows. Key capabilities include: Scrape pages into markdown, HTML, structured data, or screenshots Search the web and retrieve full page content in a single call Interact with dynamic websites using AI prompts or code Get started with Firecrawl on the Vercel Marketplace. Read more

Pull anomaly alert details using the Vercel CLI

May 21, 2026

You can now access anomaly alerts and their details directly through the Vercel CLI. With the vercel alerts command, you can list all alerts for a team or given project. For each alert, you can view the start time, the type of alert, and whether or not the alert is still active. With the --ai option, the AI investigation results appear alongside each alert. You and your agent can act on alerts without leaving the terminal. Available on Observability Plus. Learn more about vercel alerts in the CLI documentation. Read more

Qwen 3.7 Max now available on Vercel AI Gateway

May 21, 2026

Qwen 3.7 Max from Alibaba is now available on Vercel AI Gateway. The model is designed as an agent foundation, with capabilities spanning coding, office workflow automation, and long-horizon autonomous execution. Qwen 3.7 Max shows improvements in frontend prototyping and complex multi-file engineering. The model supports office and productivity tasks through multi-agent orchestration and sustains coherent reasoning across long-horizon tool-calling sessions. To use Qwen 3.7 Max, set model to alibaba/qwen-3.7-max in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Configure weighted traffic splits for Vercel Flags from the Vercel CLI

May 21, 2026

You can now configure weighted traffic splits for Vercel Flags with the new vercel flags split command in the Vercel CLI. This allows you to send a percentage of traffic to one variant and the rest to another. Run the command interactively, or pass the environment, bucketing attribute, and variant weights as flags: Update to the latest version of the Vercel CLI and read the documentation to get started. Read more

Chat SDK now includes AI SDK tools

May 20, 2026

Chat SDK now ships a built-in AI SDK toolset through the new chat/ai subpath. One createChatTools(chat) call wires Chat SDK's read and write actions into your agent. Approval by default: write tools are gated by a requireApproval option. Presets: reader, messenger, and moderator scope the toolset. Lazy loading: only the tools your preset allows are constructed. toAiMessages and its supporting types have moved to chat/ai. The previous chat re-exports are flagged @deprecated. Read the documentation to get started, or try one of our templates. Read more

Chat SDK adds message subjects and direct SDK access

May 20, 2026

You can now read the parent issue or pull request when your bot is mentioned in a Linear or GitHub comment. message.subject resolves to that parent with title, status, URL, and the full typed payload. message.subject is cached per message, so repeated access only hits the API once. It resolves to null on Slack and other chat platforms, where there's no parent resource. Direct access to platform SDKs The GitHub, Linear, and Slack adapters now expose their underlying platform SDKs. Use them to extend your bot by calling provider APIs directly. The previous .client getter remains as a @deprecated alias on the adapters. Read the documentation to get started, or explore one of our templates. Read more

Chat SDK now supports callback URLs on buttons and modals

May 20, 2026

You can now pause a Workflow run on a Chat SDK card and resume it when someone clicks a button. The same flow works for form submissions. Buttons and modals accept a new callbackUrl prop, and the event payload is sent to that endpoint. To build a card like this, create a workflow webhook and pass its URL to each button's callbackUrl prop inside your component: For the component, the form data is in the payload. callbackUrl works for buttons on most platforms with an official adapter, and for modals on Slack and Teams. Read the documentation or walkthrough guide to start building. Read more

Vercel AI Gateway plugin for WordPress

May 20, 2026

The Vercel AI Gateway plugin gives any WordPress site access to hundreds of models from 40+ providers through a single API key. Providers include Anthropic, Google, OpenAI, xAI, DeepSeek, MiniMax, Moonshot AI, and more. The plugin is implemented as a connector for the new WordPress AI Client, which requires WordPress 7.0, released today. What this enables Any AI-powered plugin works automatically. Plugins built on the WordPress AI Client pick up the connector without their own provider integrations or API keys. One key, many providers. Manage a single AI Gateway key in Settings > Connectors instead of credentials per provider. Multi-modal content generation. Text, structured JSON, image generation and editing, and video, all through the same prompt builder. Automatic fallbacks. AI features stay online during provider outages. Dynamic model discovery. New models become available without a plugin update. Unified billing and observability across providers, at provider prices. Getting started Install the Vercel AI Gateway plugin on your WordPress site Add your AI Gateway API key under Settings > Connectors To call AI Gateway directly from your own code: See the plugin documentation for

Grok Build 0.1 now available on Vercel AI Gateway

May 20, 2026

Grok Build 0.1 is now available on Vercel AI Gateway. This is a beta coding model trained for agentic coding, currently in early access, and powers the Grok Build CLI app. Reasoning effort is not configurable, and there is no non-reasoning mode. To use Grok Build 0.1, set model to xai/grok-build-0.1 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Nuxt MCP Toolkit now supports MCP apps

May 19, 2026

The Nuxt MCP Toolkit now supports MCP apps. Your agent tools can return interactive HTML responses that MCP clients like Claude and ChatGPT render inline, rather than plain-text responses. Declare a tool with the defineMcpApp macro, then read pre-hydrated data, trigger follow-up prompts, or call other tools from inside the UI with the useMcpApp composable. The toolkit bundles each Vue SFC into a self-contained HTML file at build time and serves it from your MCP endpoint. Read the Nuxt MCP Toolkit documentation to get started. Read more

Gemini 3.5 Flash on AI Gateway

May 19, 2026

Gemini 3.5 Flash is now available on Vercel AI Gateway. This model has improved coding proficiency and parallel agentic execution loops versus previous Flash versions. It also brings improvements to core reasoning, instruction following, and multi-turn coherence, with stronger performance on complex tasks and higher-quality reasoning traces in thinking mode. 3.5 Flash defaults to the medium thinking level, balancing response quality with faster, more cost-efficient generation. To use Gemini 3.5 Flash, set model to google/gemini-3.5-flash in the AI SDK. Note that temperature, topP, topK, and thinking_budget are not supported by this model. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Flat Rate CDN in Limited Beta

May 19, 2026

Flat Rate CDN is now available in Limited Beta for Pro teams. It replaces usage-based CDN pricing with a fixed monthly fee. Viral posts, unfiltered bots, or misconfigured routes can turn a normal month into a surprise bill. Flat Rate CDN makes your cost predictable. Fixed monthly price: Pay a fixed amount for Vercel CDN each month, regardless of traffic and usage. Covers spikes across Vercel CDN: Edge Requests, Fast Data Transfer, and other products are no longer metered. No overages: Your bill will never increase because of temporary usage spikes. Flat Rate CDN is a good fit for teams with unpredictable CDN bills, individual builders who don't want a usage spike to break the bank, and growing businesses that need to know their CDN costs before the month starts. Pro teams can join the waitlist for early access and to help shape the GA release. Read more

Run Claude Managed Agents with Vercel Sandbox

May 18, 2026

You can now run Claude Managed Agents with Vercel Sandbox. Claude Managed Agents handles the model, harness, tools, and session state. Self-hosting lets you bring the execution environment, so an agent's tool calls run on your existing Vercel infrastructure with your private APIs, internal services, and customer data. Each agent session runs in its own isolated Firecracker microVM, using the same infrastructure that powers 1B+ Vercel deployments with enterprise-grade security, availability, and performance. What you get: A managed agent loop from Anthropic, including the model, harness, tools, and session state A Vercel function control plane that spawns a sandbox per session Firecracker microVM isolation per session, with millisecond startup Credential brokering at the firewall, so secrets never enter the sandbox Deny-by-default egress with a domain allowlist Low-latency connectivity to your private network and cloud workloads Follow the step-by-step guide for a reference implementation and learn more in the Vercel Sandbox docs. Read more

Consolidated Commit Status now available on GitHub

May 18, 2026

Monorepos can now opt in to a single, consolidated commit status on pull requests instead of one commit status per project. For repos with many projects, teams can configure GitHub branch protection once, then manage which Vercel projects are required for merge in each project’s settings. Read more about Github Commit statuses and and enable Consolidated Commit Status from your project settings. Read more

Firewall‑mitigated traffic is free on Vercel

May 18, 2026

Vercel Firewall now waives CDN Requests and Fast Data Transfer for any traffic denied, challenged, or rate‑limited by Web Application Firewall (WAF). Vercel has always provided unlimited DDoS mitigation at no cost. Vercel WAF, included in CDN cost, gives you custom rules, managed rules, and rate limiting for bad traffic that isn't DDoS. With this change, you don't pay for requests or bandwidth that WAF denies, challenges, or rate‑limits. That means no surprise bill when a scraper hammers your product pages, a credential‑stuffing botnet hits your login route, or a bot abuses an expensive endpoint. The waiver applies automatically to every project using Vercel Firewall and no configuration is required. Learn how to implement WAF rules in the Firewall documentation. Read more

Public Web Analytics API now available

May 18, 2026

The Web Analytics API is now publicly available, with programmatic access to page views, visitors, and custom events for your Vercel projects. The API reads from the same aggregated data model that powers the Web Analytics dashboard, so numbers from the API will match what your team sees in the UI. You can use it to build your own custom reports or to embed live traffic metrics in a customer-facing admin panel. Two resources are exposed: visits for automatically tracked page views and visitors events for custom events sent with track() Count endpoints return a single total; aggregate endpoints return rows grouped by time, route, country, referrer, device, feature flag, or custom event data. Filters use OData syntax, with by to group results. For example, to count visitors of a specific page, call visits/count with a requestPath filter: For raw, unaggregated event streams, forward analytics data to your own infrastructure with Drains. Read the Web Analytics API guide for authentication setup, full filter syntax, and response schemas. Read more

Use native curl syntax with Vercel CLI

May 15, 2026

You can now use native curl syntax with the Vercel CLI. The vercel curl command accepts full URLs, bare hostnames, and the --url flag, and uses your Vercel auth to bypass Deployment Protection. If you've linked a project, you can also pass just a path: Update to the latest Vercel CLI version and run vercel curl to get started. Learn more in the Vercel CLI documentation. Read more

Sort providers by cost, latency, or throughput on AI Gateway

May 15, 2026

You can now sort the providers behind a model by cost, time to first token (TTFT), or throughput (TPS) in AI Gateway. The default provider order blends provider reliability, quality of model output, cost, and speed of response. You can now use sort for explicit control over ranking criteria. For models with many providers and noticeable cost or speed variation, you can use sort to optimize on your dimension of choice. Ranking is computed at request time, so newly added providers, price changes, and shifts in observed latency or throughput flow through automatically without any code changes. Set sort on providerOptions.gateway to one of the three values: Value Description Direction When to use 'cost' Sort by the provider's listed input price per million tokens Lowest price first High-volume, cost-sensitive work 'ttft' Sort by median time to first token, in ms Lowest latency first Latency-sensitive workloads where response speed matters 'tps' Sort by median tokens per second throughput Highest first Long-output generation where total response time matters most Basic usage Use sort to ensure optimizing for your metric of choice. In this example, AI Gateway has over five providers for

Protected Source Maps: Ship browser source maps securely

May 14, 2026

You can now restrict access to production source maps with Protected Source Maps, which puts browser .map files behind Vercel Authentication. Your team can fetch them; everyone else gets a 404. Source maps are how you debug minified production code. They give you readable stack traces and your original source code, with the real filenames and line numbers intact. New projects have Protected Source Maps enabled by default. Existing projects can opt in from Settings → Deployment Protection, with no redeploy needed. Learn more in the documentation. Read more

Trusted Sources for Deployment Protection

May 13, 2026

Trusted Sources lets protected deployments accept short-lived identity tokens (OIDC) from Vercel projects and external services you authorize, so you no longer have to share a long-lived Protection Bypass for Automation secret. Trusted Sources is the recommended approach, but Protection Bypass for Automation continues to work Callers attach an OIDC token in the x-vercel-trusted-oidc-idp-token header. Vercel then verifies the signature, checks the claims you configured, and confirms the environment matches the rule. Authorize Vercel projects By default, the Vercel OIDC token for a project can call its own deployments. To authorize another project in the same team, add it to Trusted Sources. Self-access and cross-project rules are both customizable with from/to environment pairs. To authenticate a request from a project, forward its Vercel OIDC token: Authorize external services Any custom OIDC provider can be authorized as a trusted external service, such as GitHub Actions, or a Vercel project in another team. Read the documentation to learn more. Read more

Fast mode for Opus 4.7 available on AI Gateway

May 12, 2026

Fast mode for Claude Opus 4.7 is now available on AI Gateway in research preview. Fast mode delivers ~2.5x faster output token generation with full Opus 4.7 intelligence. This is an early, experimental feature. To enable fast mode, pass speed: 'fast' in the anthropic provider options with anthropic/claude-opus-4.7. You can use fast mode with Claude Code via AI Gateway by setting the CLAUDE_CODE_SKIP_FAST_MODE_ORG_CHECK and CLAUDE_CODE_ENABLE_OPUS_4_7_FAST_MODE variables in your shell configuration file or in ~/.claude/settings.json. Fast mode is priced at 6x standard Opus rates. Standard Fast Mode Input: $5 / 1M tokens Output: $25 / 1M tokens Input: $30 / 1M tokens Output: $150 / 1M tokens All standard pricing multipliers (e.g., prompt caching) apply on top of these rates. Read more

AI Gateway production index

May 12, 2026

Ask which AI model is best, and the answer changes before the ink dries. That's what happens in an industry where new models are released weekly. Every benchmark measures a different race, and every race crowns its own winner, but Vercel has a unique view of the industry through production workloads. AI Gateway serves tens of trillions of tokens across hundreds of models through real applications and agents. What we're seeing: Anthropic leads in spend despite a higher unit price, Google leads in volume OSS models are gaining traction, but there is no loyalty to specific labs OpenAI spend share is growing quickly after recent model updates High-volume workloads route across 30+ distinct models on average Agentic workloads carry 59% of all token volume (up 2x over 6 months) This report is built on data from seven months of production traffic from AI Gateway, with usage from over 200K+ unique teams. Anthropic leads in spend; Google leads in volume Cost and volume rankings disagree because they measure two different workloads, even for the same customer. By spend in April 2026, Anthropic took 61%, Google 21%, and OpenAI 12%. By token volume, the picture flipped. 38% of April traffic th

Manage Vercel Firewall in the CLI

May 12, 2026

You can now manage your project's Vercel Firewall configuration through the Vercel CLI. Use the new vercel firewall command to manage custom rules, IP blocks, system bypasses, Attack Mode, and system mitigations from the terminal. The Vercel Firewall skill also helps agents by detailing best practices for rolling out new rules safely. Update Vercel CLI and run vercel firewall to get started. Learn more about the Vercel Firewall CLI commands. Read more

Node.js 26.x now available on Vercel Sandboxes

May 12, 2026

Vercel Sandbox now supports Node.js version 26. To run a Sandbox with Node.js 26, upgrade @vercel/sandbox to 1.10.2 or later, or to 2.0.0-beta.19 or later if you're using v2 and set the runtime property to node26: Get started today and learn more in the documentation. Read more

Create Vercel Firewall rules with natural language

May 12, 2026

Vercel Firewall now lets you create WAF custom rules using natural language. Describe the behavior you need and the dashboard will generate the rule. Visit the firewall custom rules page to try creating a rate-limiting rule: Or use the Vercel CLI: WAF custom rules let you control traffic to your site by logging, blocking, challenging, rate limiting, or redirecting requests based on conditions like IP address, path, country, user agent, and more. For example, you can: Log all requests to /api/webhook with a missing authorization header Block all requests to /wp-admin Challenge all traffic to /checkout that doesn't come from the US Generate your first rule or learn more in the documentation. Read more

Automate progressive rollouts with Vercel Flags

May 11, 2026

You can now use Vercel Flags to roll out a feature to a growing percentage of users on a schedule, with progressive rollouts. Unlike weighted splits, which hold a fixed distribution (for example, 50/50) for experiments, a progressive rollout follows a predefined schedule that gradually shifts the traffic percentage to the new variant. Each stage has a target percentage and a duration. Exposing a change in stages lets you catch a regression on a small slice of users before it hits everyone. Progressive rollouts are now available in the dashboard and through the new vercel flags rollout CLI command. Learn more in the Vercel Flags documentation. Read more

Vercel Sandbox firewall now supports request proxying and filtering

May 11, 2026

The Vercel Sandbox firewall now supports forwarding specific HTTP requests to a proxy you control. You can also use matchers to filter forwarding and credentials brokering to only the requests that need it. Requests proxying You can now route outbound sandbox traffic through your own proxy for logging, debugging, or transforming requests and responses. Set a forwardURL on any allowed domain, and the firewall will forward matching HTTPS requests to your server. The proxy receives the original request along with additional headers to identify the source: vercel-forwarded-host: The original request's SNI vercel-forwarded-scheme: The original request's scheme vercel-forwarded-port: The original request's port vercel-sandbox-oidc-token: A Vercel-issued OIDC token that the proxy can use to authenticate the request and identity the source team / project / sandbox. Learn more about it in the docs Filtering Additionally, you can now use matchers to filter request forwarding or credentials brokering to requests matching a specific path, method, query string, or headers. This gives you fine-grained control over which requests get transformed; for example, only forwarding POST requests to a sp

How Superset built the IDE for AI agents on Vercel

May 10, 2026

Superset on Vercel 1,000–1,400 deployments per week ~600 preview deployments per day ~30 second average build time 57–64% week-over-week DAU growth Software development with AI started as a single engineer chatting with a single agent about a local repo. Today, developers direct fleets of agents in the cloud, but traditional tools were built for the old shape of the job: IDEs, terminals, and review systems designed for one developer moving one ticket at a time. Co-founders Kiet Ho, Satya Patel, and Avi Peltz, all former CTOs at YC-backed companies, built Superset as the IDE for multi-agent development. It runs up to 10 coding agents in parallel, each in its own isolated workspace. Developers use it to direct teams of agents generating code across multiple branches simultaneously. Running a team of agents in parallel changes what the platform underneath has to do. The product Superset offers its users only feels parallel because nothing on the platform forces the work to wait. If any layer slows down, even briefly, the parallelism on top collapses with it. Parallel agents need parallel infrastructure This workflow has a dependency that's invisible from the product surface. Every age

Chat SDK adds Messenger adapter support

May 8, 2026

Chat SDK now supports Messenger, extending its single-codebase approach to Slack, Discord, GitHub, Teams, and Telegram with the new Messenger adapter. Teams can build bots that support messages, reactions, attachments, postback buttons, and direct conversations, with display names fetched automatically from user profiles. Try the Messenger adapter today: The adapter does not support message history, editing, or deletion. Certain components, like tables and fields, render as text fallbacks. Streaming is handled by accumulating and then sending the final message. Read the documentation to get started, browse the directory, or build your own adapter. Special thanks to @mitkodkn, whose community contribution in PR #461 laid the groundwork for this adapter. Read more

Chat SDK adds web adapter support

May 8, 2026

You can now build chat UIs that connect to Chat SDK with the new web adapter. Build in-product assistants, support agents, or any other browser-based chat experience. Define the bot on your server: Then stream replies to the browser with a preconfigured @ai-sdk/react useChat hook: Read the documentation to get started, browse the directory, or build your own adapter. Read more

Chat SDK now supports conversation history

May 8, 2026

Chat SDK now supports cross-platform conversation history through the new transcripts and identity options. User transcripts persist across every platform adapter, allowing the same user to keep their message history wherever they message your bot. bot.transcripts exposes four methods, backed by one of the official state adapters: append: persist an inbound message or a bot reply list: return entries chronologically with filters count: total entries stored for a user delete: wipe every entry for a user Read the documentation to get started, or try one of the templates. Read more

Next.js May 2026 security release

May 7, 2026

Summary We have shipped a coordinated security release for Next.js addressing 13 advisories across denial of service, middleware and proxy bypass, server-side request forgery, cache poisoning, and cross-site scripting. One advisory addresses an upstream React Server Components vulnerability tracked as CVE-2026-23870. Recommended actions Patched versions are available for both React and Next.js, and all affected users should upgrade immediately. Impact The release addresses the following advisories: Middleware and proxy bypass Affects applications that rely on middleware.js or proxy.js for authorization. High: Auth bypass via App Router segment-prefetch URL High: App Router segment-prefetch bypass, incomplete fix follow-up High: Pages Router i18n default-locale path bypasses proxy authorization High: Bypass via dynamic route parameter injection Low: Middleware redirects can be cache-poisoned Denial of service Affects applications using Server Functions, Partial Prerendering with Cache Components, or the Image Optimization API. High: DoS in React Server Components (tracked upstream as CVE-2026-23870) High: DoS via connection exhaustion in applications using Cache Components Moderate:

Vercel Flags now supports JSON values

May 7, 2026

You can now store JSON values in Vercel Flags, extending the existing support for boolean, string, and number values. This allows you to collapse what used to take several related flags into a single feature flag. For example, to A/B test how a different model performs, you can now define a single model flag. This allows you to manage one flag that serves the full object rather than managing ai_model, ai_temperature, and ai_max_tokens separately: Use Vercel Flags to progressively route traffic to a new model, A/B test, or quickly switch models in case a provider is having issues. Try it out or learn more about Vercel Flags. Read more

Auto-add Git committers to your team

May 6, 2026

Pro teams can now choose how Git committers to private repositories are added to their Vercel team. Auto Approval: non-team committers with Vercel accounts are automatically added to your team and their deployments proceed immediately. Added members count toward your team seats at standard Pro pricing. Manual Approval: deployments are blocked until an owner approves the new member. Choose your approval preference in team settings. Learn more about collaboration settings and troubleshooting project collaboration. Read more

Secure Marketplace credentials with Production-only access

May 6, 2026

You can now secure native integration resources by restricting where they can be used. Setting a resource to Production only removes non-production access and protects credentials as sensitive environment variables. This makes it so secret values or no longer readable from the dashboard or CLI From the integration resource Settings, select Allowed Environments → Production only and save. We recommend that you rotate the secrets of the integration resource after saving. Once applied: Development and Preview connections are removed New non-production connections are blocked Connections without a Production target are disconnected Credentials are protected and no longer readable Reverting this setting requires Owner permissions for your Vercel team. Owners can re-enable Development and Preview environment from Settings and reconnect projects if needed. You may be prompted to reauthenticate with an MFA challenge depending on your settings. Learn more in the Vercel documentation. Read more

Query observability metrics using the Vercel CLI

May 5, 2026

You can now access Observability Plus metrics in the Vercel CLI. Query observability data for any Vercel team or project using the new vercel metrics command. Coding agents can also leverage this new command to better analyze the performance, reliability, or security of applications on Vercel, as well as debug issues. This feature is available in public beta for all teams with Observability Plus. Learn more about vercel metrics. Read more

How KIKO Milano scales for Black Friday

May 5, 2026

KIKO Milano on Vercel: Eliminated 3 weeks of Black Friday infrastructure prep 75% decrease in app build times Went from minimal releases to deploying multiple times per day KIKO Milano’s ecommerce team used to treat peak traffic as an operations project. Weeks before Black Friday, they had to manually scale AWS infrastructure and adjust application configuration, knowing that if demand exceeded forecasts, their site could slow down or even break, costing them real revenue. After migrating to Vercel, they removed manual prep from their playbook. Instead of provisioning infrastructure, their team now ships campaigns and tests daily. The Black Friday problem: manual scaling and stressful failure modes Before Vercel, KIKO ran their ecommerce app on AWS EC2 instances sized for normal traffic, then scaled by hand ahead of peak periods. This was their previous process leading into Black Friday: Manual scaling window: Black Friday prep began 2–3 weeks ahead of the event, then had to be unwound after the traffic spike passed. Infrastructure configuration: The infra team manually adjusted AWS EC2 capacity to support expected traffic. Application configuration: Scaling also required applicati

How General Intelligence used agents to build an agent platform on Vercel

May 4, 2026

General Intelligence on Vercel 8-person team (5 engineers) shipping 10 PRs and 70+ commits per engineer, per day 4,000+ preview branches with ~100 parallel app versions running at any moment 90% of SRE work automated through Vercel and their own agent (Cofounder) Cofounder launches with a managed Vercel account for every customer General Intelligence is building a platform that lets any founder run a company entirely with AI agents. Their vision is the one-person, billion-dollar company where every department is driven by agents. Their flagship product, Cofounder, gives founders a full team of agents that cover engineering, marketing, SEO, finance, sales, customer support, and operations. General Intelligence is an 8-person company with 5 engineers. To ship a platform that lets their customers run agentic companies, they had to operate as one themselves. They wanted to use Cofounder's CTO agent to build out other agentic business functions, and their developers quickly learned that was an infrastructure problem, not an agent problem. For a complex, multi-tenant platform, the requirement was total programmatic control: every action a human could take on the underlying cloud platform

Introducing deepsec: The security harness for finding vulnerabilities in your codebase

May 4, 2026

Today we’re open sourcing deepsec: a security harness powered by coding agents. It runs on your own infrastructure and surfaces hard-to-find issues in large codebases. You can run deepsec on your laptop without setting up a cloud service for privileged source code access. For inference, you can use your existing Claude or Codex subscription without any additional setup. Scanning large repos can take multiple days on a single machine. To run research jobs in parallel, deepsec supports optional fanout to Vercel Sandboxes for remote execution. Scans on Vercel’s codebases routinely scale up to 1,000+ concurrent sandboxes. Architecture At its core, deepsec uses claude and codex to perform tailored investigation of a codebase using Opus 4.7 at max effort and GPT 5.5 at xhigh reasoning. Scans start with static analysis to identify security-sensitive files, then coding agents investigate each candidate, tracing data flows, checking for mitigations, and producing actionable findings with severity ratings. Here is the workflow: Scan: It starts by performing a regex-only scan of all files for security-sensitive areas that subsequent steps will focus on. Investigate: Agents investigate each fi

Postgres connections now work through Sandbox firewall

May 1, 2026

Vercel Sandbox can now connect to hosted Postgres databases, including Neon, Supabase, AWS RDS, Nile, and Prisma Postgres. To enable a connection, add the database host to your Sandbox's allowed domains. Background When SNI based filtering is used with Vercel Sandbox, the sandbox firewall restricts outbound network access by checking the domain name during a connection's TLS handshake. This works seamlessly for HTTPS traffic, where the domain is visible at the start of the connection. Postgres, however, negotiates TLS differently. A Postgres client first opens a plain TCP connection and then upgrades to TLS. Because the domain isn't available when the firewall first needs it, Postgres connections through a standard domain-restricted Sandbox would fail. What changed The Sandbox firewall now adjusts for the Postgres TLS negotiation flow. It detects the protocol's startup sequence, waits for the TLS upgrade, and then applies your domain policy before forwarding the connection to the database. No changes are needed to your code or database configuration. Connecting to hosted database Here's a full example: create a Sandbox, install a Postgres client, lock down the network to only the d

How GitBook serves 30,000 sites with sub

May 1, 2026

GitBook on Vercel 30,000 documentation sites hosted on a single Vercel deployment 120 million monthly page views served from the edge 40,000 cache invalidations processed daily, each resolved in under 300ms 41% of all traffic now comes from AI crawlers and automated systems GitBook hosts 30,000 documentation sites on Vercel, serving 120 million page views every month. Companies like n8n, Nvidia, and Zoom trust the platform to keep their docs fast and current. For modern engineering teams and coding agents, documentation is as critical as the production code it describes, and GitBook sits at the center of that expectation. GitBook's publishing model mirrors how teams ship software: propose changes, review, and merge. With 30,000 independently managed sites each on its own update cadence, keeping content fast and fresh was a complicated engineering problem. Before migrating to Vercel, site editors would hit merge, visit the published site to validate, and see old version of the content. "One of our customers planned a large feature release, and on launch day, the docs lagged behind the rest of the product," says Steven Hall, Head of Engineering at GitBook. "That was the moment we rea

Grok 4.3 on AI Gateway

April 30, 2026

Grok 4.3 is now available on Vercel AI Gateway. The model has a 1M token context window and improvements in accuracy, tool calling, and instruction following. To use Grok 4.3, set model to xai/grok-4.3 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Custom tags available in beta on Vercel Sandbox

April 30, 2026

As teams scale isolated environments for AI agents, code generation, or dev workflows, keeping track of which sandbox belongs to whom, and why, becomes critical. Custom tags allow you to organize, filter, and manage Vercel Sandboxes at scale. Each sandbox supports up to five tags. Organize by environment, team, or customer Tags are flexible by design. Use them to separate staging from production, attribute usage to specific teams, or isolate sandboxes per customer in multi-tenant platforms: Update tags as context changes Promote a sandbox from staging to production, reassign ownership, or mark it for cleanup without recreating it: Easily track your sandboxes Filter sandboxes by any tag to quickly surface the ones that matter. This is useful for dashboards, cleanup scripts, or routing logic that needs to find all sandboxes matching a specific environment or team: Use Cases AI agents at scale: Tag sandboxes by session, user, or agent run to track which execution environment belongs to which workflow. Multi-tenant platforms: Isolate and filter sandboxes per customer or workspace, making billing attribution and cleanup straightforward. Team-level visibility: Attribute sandbox usage to

Vercel now supports Pro plan in Stripe Projects

April 29, 2026

You can now sign up for or upgrade to a Vercel Pro plan directly from Stripe Projects using shared payment tokens (SPTs). Agents and developers can manage plan changes programmatically from the Stripe CLI, without leaving their workflow. What’s new Provision or upgrade to Vercel Pro directly from the Stripe CLI Support for both upgrade and downgrade flows Powered by shared payment tokens for secure, streamlined billing This builds on our Stripe Projects launch in developer preview by enabling end-to-end provisioning and billing in one place. Instead of switching between dashboards, you can now handle infrastructure setup and plan management directly from the terminal. Getting started If you’re already using Stripe Projects and have set up billing via stripe projects billing add , you can upgrade your Vercel plan from the CLI simply by running stripe projects add vercel/pro If you are new to Stripe Projects, Install the plugin and initialize your project: Read more

Native Deployment Checks are now available

April 28, 2026

You can now run lint and typecheck on every Vercel deployment, in parallel with the build. Native Deployment Checks are available to every team and join your existing Deployment Checks alongside GitHub and Marketplace integrations. Once added from your project's Build and Deployment settings, Vercel runs the matching script from your package.json on each deployment, and skips the check if no matching script exists. You can mark a check as required to hold the deployment from production until it passes, and choose which environments each check runs on. When a Native Deployment Check fails on a pull request, Vercel Agent investigates the failure and suggests a fix you can review and merge. Read more

2026 Vercel AI Accelerator recap

April 28, 2026

On April 16th, 39 teams took the stage to pitch investors at Demo Day. During the prior six weeks, founders worked shoulder-to-shoulder with the Vercel team, our partners, and industry leaders to shape their ideas into the next generation of AI applications. Six weeks with the 2026 cohort Teams from around the world joined the 2026 cohort, building agents, developer tools, consumer apps, and vertical AI for finance, security, healthcare, and robotics. Each week, the cohort joined two sessions: a hands-on technical workshop and a fireside chat with an external speaker. The workshops covered the technical side of building AI products, from agents and models to deployment and scale. The fireside chats brought in industry leaders, Vercel experts, and investors. Speakers included Peter Steinberger (OpenAI/OpenClaw) and the Windsurf team, with topics ranging from the state of AI to pricing, legal, sales, marketing, and fundraising. Builder day Halfway through the program, the cohort came together with the broader Vercel startup ecosystem for Builder Day at Vercel's office. The day combined technical workshops, hands-on office hours with engineers from accelerator partners like AWS and An

Hobby projects now default to 30-day deployment retention

April 27, 2026

Starting April 29th, the maximum retention policy for Hobby plans will be capped at 30 days. Deployments outside your retention window will be automatically removed. This excludes your 10 most recent production deployments and any aliased deployments, which continue to be preserved regardless of retention settings. Pro and Enterprise plans are not affected. Learn more about Deployment Retention. Read more

GPT 5.5 on AI Gateway

April 24, 2026

GPT-5.5 is now available on Vercel AI Gateway. There are 2 variants: GPT-5.5 and GPT-5.5 Pro. Both models are tuned for long-running agentic work across coding, computer use, knowledge work, and scientific research, and are more token-efficient than the previous generation. GPT-5.5 is stronger at agentic coding and long-horizon work where the model needs to hold context across a large system and carry changes through the surrounding codebase. Paired with computer-use skills, it can operate real software and turn raw material into documents, spreadsheets, or slide presentations. GPT-5.5 Pro is built for demanding, multi-step work where response quality matters more than latency. Early testing shows gains in business, legal, education, data science, and technical research workflows that involve critiquing work over multiple passes and stress-testing arguments. To use GPT-5.5, set model to openai/gpt-5.5 or openai/gpt-5.5-pro in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Ke

Deepseek V4 on AI Gateway

April 23, 2026

DeepSeek V4 is now available on Vercel AI Gateway. There are 2 model variants: DeepSeek V4 Pro and DeepSeek V4 Flash. A 1M token context window is the default across both models. DeepSeek V4 Pro focuses on agentic coding, formal mathematical reasoning, and long-horizon workflows. It handles feature development, bug fixing, and refactoring across stacks, with tool use that works across harnesses like MCP workflows and agent frameworks. It also writes clear, well-structured long-form documents. DeepSeek V4 Flash performs close to V4 Pro on reasoning and holds up on simpler agent tasks, with a smaller parameter size for faster responses and lower API cost. It's a good fit for high-volume workloads and latency-sensitive use cases. To use DeepSeek V4, set model to deepseek/deepseek-v4-pro or deepseek/deepseek-v4-flash in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gat

GPT Image 2 on AI Gateway

April 21, 2026

GPT Image 2 is now available on Vercel AI Gateway. OpenAI's newest image model supports detailed instruction following, accurate placement and relationships between objects, and rendering of dense text across multiple aspect ratios. The model can render fine-grained elements including small text, iconography, UI elements, dense compositions, and subtle stylistic constraints, at up to 2K resolution. Non-English text is also supported and reads coherently. GPT Image 2 can produce photos, cinematic stills, pixel art, manga, and other distinct visual styles, with consistency in texture, lighting, composition, and detail. This suits workflows like game prototyping, storyboarding, marketing creative, and medium-specific asset generation. To use GPT Image 2, set model to openai/gpt-image-2 in the AI SDK, or try it directly in our model playground. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, vi

Kimi K2.6 on AI Gateway

April 20, 2026

Kimi K2.6 from Moonshot AI is now available on Vercel AI Gateway. The model focuses on long-horizon coding tasks, with generalization across languages such as Rust, Go, and Python and across front-end, devops, and performance optimization work. K2.6 can turn simple prompts into complete front-end interfaces with structured layouts. For autonomous, proactive agents that run continuously across multiple applications, K2.6 improves on API interpretation, long-running stability, and safety awareness during extended research tasks. To use Kimi K2.6, set model to moonshotai/kimi-k2.6 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Deployment retention policies now preserve active branch deployments

April 17, 2026

Retention policies no longer delete the latest preview deployment for branches with open or unmerged pull requests. Previously, deployments for active branches could be removed if they exceeded the configured retention window. This means you can safely use shorter retention windows without risking losing active preview deployments. This change applies to all plans. Your 10 most recent production deployments and any aliased deployments continue to be preserved regardless of retention settings. Learn more about Deployment Retention. Read more

How Zo Computer improved AI reliability 20x on Vercel

April 17, 2026

Zo Computer on Vercel 20x reduction in retry rate (7.5% → 0.34%) 99.93% chat success rate (up from 98%) P99 latency cut 38% (131s → 81s) New models added in less than 1 minute Every company has servers that store data, run services, and do work around the clock. Consumers just have apps. Rob Cheung, co-founder of Zo Computer, is closing that gap. Zo is a personal AI cloud: your own servers and data that power an always-on agent. "Cloud is one of the best computing models of all time, and consumers have zero direct access because it's so complicated," explained Rob Cheung, co-founder and CEO of Zo. "Now, with AI, it's finally possible for all of us to have cloud computers." Zo is a full computing environment, not just a chatbot. Rob laughs about his mom running servers and databases without knowing it. People use Zo to manage small businesses, do research, organize finances, and track health data. The 8-person company is two and a half years old and they have an ambitious goal: to onboard one million new users to personal cloud computing in 2026. That means millions of AI model calls every day, and when Zo users text their agent like a friend, they expect the same responsiveness. De

Claude Opus 4.7 on AI Gateway

April 16, 2026

Claude Opus 4.7 from Anthropic is now available on Vercel AI Gateway. Opus 4.7 is optimized for long-running, asynchronous agents and handles complex, multi-step tasks with reliable agentic execution. The model shows gains on knowledge-worker tasks, particularly where it needs to visually verify its own outputs. Opus 4.7 is also stronger at programmatic tool-calling with image-processing libraries to analyze charts and figures, including pixel-level data transcription. It has high-resolution image support, which is useful for computer use, screenshot understanding, and document analysis workflows. Opus 4.7 now has improved memory, with agents that maintain structured memory store across turns seeing more reliable recall and fewer dropped facts without additional prompting. To use Claude Opus 4.7 set model to anthropic/claude-opus-4.7 in the AI SDK. You can also try a new effort level: 'xhigh'. Opus 4.7 also introduces the task budgets feature. Task budgets let you set a total token budget for an agentic turn via taskBudget. The model sees a countdown of remaining tokens, which it uses to prioritize work, plan ahead, and wind down gracefully as the budget is consumed. Thinking conte

A new programming model for durable execution

April 16, 2026

The gap between prototypes and production-ready systems is huge. Code that's trivial to run locally falls apart the moment it needs to handle failures, restarts, and real traffic. Framework defined infrastructure solved this for web applications. When you deploy, Vercel infers the right configuration from the app itself. Workflows extends that model to long-running systems. Instead of managing a separate codebase for orchestration, durable workflows are an extension of your application code. Since launching in beta in October 2025, Workflows has processed over 100 million runs and over 500 million steps across more than 1,500 customers, with more than 200K npm downloads every week. Today, Vercel Workflows is generally available. Built for agents, backends, and long-running workloads Workflows is built for any workload that doesn’t fit in a single request. Agents: Deep integration with the AI SDK enables infinitely long running durable agents that can maintain state, tools, and handle external events or interruptions. AI SDK v7 is taking this further with WorkflowAgent. Backends: Workflow SDK proved this programming model in TypeScript codebases, and we're bringing it to a new langu

Vercel Flags is now generally available

April 16, 2026

Vercel Flags is now generally available. Vercel Flags is a feature flag provider built into the Vercel platform. Create and manage feature flags with targeting rules, user segments, and environment controls directly in the Vercel Dashboard. The Flags SDK provides a framework-native way to define and use these flags within Next.js and SvelteKit applications, integrating directly with your existing codebase: Once you define a flag, you can use them within your application in a few lines of code: For teams using other frameworks or custom backends, the Vercel Flags adapter supports the OpenFeature standard, allowing you to plug Vercel Flags into their provider agnostic SDK. Try it out or learn more about Vercel Flags. Read more

Seedance 2.0 Video Generation on AI Gateway

April 15, 2026

You can now access Bytedance's latest state-of-the-art video generation model, Seedance 2.0, via AI Gateway with no other provider accounts required. Seedance 2.0 is available on AI Gateway in two variants: Standard and Fast. Both share the same capabilities. Standard produces the highest quality output, while Fast prioritizes generation speed and lower cost. Seedance 2.0 is strong at maintaining motion stability and fine detail across frames, producing consistent output even in complex scenes with facial expressions and physical interactions. The model also generates synchronized audio natively, with support for speech in multiple languages and dialects. Beyond text-to-video and image-to-video, Seedance 2.0 adds multimodal reference-to-video, letting you combine image, video, and audio inputs as reference material in a single generation. It also supports video editing and video extension, along with professional camera movements, multi-shot composition, and in-video text rendering. To use this model, set model to bytedance/seedance-2.0 or bytedance/seedance-2.0-fast in the AI SDK or try it out in the AI Gateway Playground. Text to Video Generate video from a text prompt. Describe

Reduced pricing for Turbo build machines

April 15, 2026

Vercel is reducing the price of Turbo build machines by 16%. All builds are now priced at $0.0035 per CPU per minute. With this new model: Turbo machines, with 30 CPUs, are now $0.105 per minute (previously $0.126) Enhanced machines, with 8 CPUs, continue to be priced at $0.028 per minute Standard machines, with 4 CPUs, continue to be priced at $0.014 per minute This change will begin rolling out on April 27, and will appear on invoices for the current billing cycle as "Build CPU Minutes". Learn more about build machine pricing or monitor your Builds usage from Project Usage. Read more

Logs filtering for Vercel Workflows now available

April 14, 2026

Workflow run log filtering is now supported on Vercel, making it easy to view all logs associated with a workflow run in one place instead of piecing them together across individual requests. You can use the “View Logs” button from the workflow run details page to jump directly into the Logs tab. From there, filter logs using the Workflow Run ID and Workflow Step ID to quickly find logs for specific runs or steps. Workflow runs on Vercel already provide run-level observability, including step progression, payloads, outputs, and performance metrics. With this addition, you can now also access all related logs directly from the familiar Vercel Logs dashboard. Learn more about the Workflow SDK. Read more

Elastic Build Machines is now GA

April 14, 2026

Elastic build machines, released in beta on March 24, are now generally available for all Pro and Enterprise customers, and are now the default for all new Pro teams. Rather than a one-size-fits-all approach, Vercel evaluates each project individually and assigns the right machine for its actual needs balancing speed and cost. Over 400 teams and 6,000 projects have enabled Elastic as their default build machine. During the beta, approximately 80% of projects have reduced their costs by switching to smaller build machine while maintaining their build speeds. The remaining 20% of projects have been auto upgraded to machines with more CPUs and memory improving their build speed. Enable elastic builds in your team settings or project settings, or read the builds documentation. Read more

Anomaly alerts are now generally available

April 13, 2026

Anomaly alerts are now generally available. Teams using Observability Plus receive alerts when anomalies are detected in their applications to help quickly identify, investigate, and resolve unexpected behavior. Alerts help monitor your app in real-time by surfacing unexpected changes in usage or error patterns: Usage anomalies: unusual patterns in your application metrics, such as edge requests or function duration. Error anomalies: abnormal error patterns, such as sudden spikes in 5XX responses on a specific route. Once an anomaly is detected, Vercel Agent can automatically investigate the issue, identify the likely root cause, analyze the impact, and suggest next steps for remediation. View alerts directly in your dashboard, or subscribe via email, Slack, or webhooks to get notified wherever your team works. You can also customize what alerts you receive using alert rules. This feature is available for all teams with Observability Plus at no additional cost. Try it out or learn more about Alerts. Read more

Copy-to-Prompt instructions now available for Flags

April 13, 2026

The feature flags details page now includes copy-to-prompt instructions in the instructions pane. You or your agent can install the Flags SDK, link the project using the Vercel CLI, and add the required flag definitions to the code base. Teams that prefer manual configuration can still copy the exact flag definitions directly from the instructions pane. Learn more about Vercel Flags. Read more

Anomaly alert configuration now available

April 10, 2026

You can now granularly configure anomaly alerts to define exactly which unexpected spikes and errors matter to your application. Alert rules give you detection-level control, allowing you to customize which projects, alert types, metrics, HTTP status codes, and specific routes you monitor for anomalies. For the anomalies you choose to track, Vercel automatically investigates your logs and metrics to identify the root cause, routing those findings to distinct destinations like dedicated Slack channels or individual emails. When you configure a rule to silence a specific pattern, anomaly detection skips that traffic entirely, preventing anomalies from appearing in your dashboard and stopping notifications before they trigger. This feature is available for teams using Observability Plus at no additional cost. Try it out or learn more about alert rules. Read more

Zero-configuration Django support

April 9, 2026

Django, one of the most popular high-level Python web frameworks, for rapid development, is now supported with zero-configuration. You can now instantly deploy Django full-stack apps or APIs on Vercel. Vercel now recognizes and deeply understands the specifications of Django applications, removing the need for redirects in vercel.json or using the /api folder. All applications on Vercel use Fluid compute with Active CPU pricing by default. Static files will be served by the Vercel CDN. Deploy Django on Vercel or visit the Django on Vercel documentation Example Django app CLI entry point Project settings URL routing table Request handlers WSGI entry point Read more

Agentic Infrastructure

April 9, 2026

Every generation of software eventually demands a new generation of infrastructure. First, we configured servers by hand. Next, the cloud turned infrastructure into APIs. Then, a more important shift: infrastructure derived from the application itself. LLMs and coding agents are driving the next transition, and it's happening fast. In the three months leading up to this post’s publication, weekly deployments on Vercel doubled, with agents driving the growth. At the time, over 30% of deployments were initiated by coding agents, a 1000% increase from six months earlier. As of September 2026, that share has risen to more than 50%. Software is now agentic Agents are building, testing, and shipping AI-native software, and they're doing it at a velocity that breaks traditional operations. Vercel projects deployed by coding agents are 20 times more likely to call AI inference providers than those deployed by humans. Agents are writing software that uses AI, and agents are building agents. As the final actor shifts from human to machine, infrastructure has to adapt again. It has to work for software that acts on behalf of users, writes itself, and increasingly needs to understand its own b

Vercel Sandbox now supports up to 32 vCPU + 64 GB RAM configurations

April 8, 2026

Vercel Sandbox now supports creating sandboxes with up to 32 vCPUs and 64 GB of RAM for Enterprise customers. This enables running large, resource-intensive applications that are CPU-bound or require a large amount of memory. Get started by setting the resources.vcpus option in the SDK: Or using the --vcpus option in the CLI: Learn more about Sandbox in the docs. Read more

Zero Data Retention on AI Gateway

April 8, 2026

Building with multiple AI models means wrestling with fragmented data policies. With many different model providers, it's not just fragmented, it's just too much time spent on the wrong things. You have to read through different terms of service, track which providers comply with your security requirements, and hope developers remember to configure opt-outs correctly on every request. What should be a straightforward policy becomes a manual, error-prone process because many providers do not offer Zero Data Retention (ZDR) by default. AI Gateway changes this by handling the negotiation and enforcement for you. Instead of managing policies provider by provider, you get the freedom to just build. AI Gateway ensures your data requirements are met automatically by only routing to providers where we have negotiated Zero Data Retention agreements. Models from OpenAI, Anthropic, Google, and more have ZDR providers available. Today, we are expanding AI Gateway's compliance capabilities with team-wide Zero Data Retention (ZDR), letting you enforce strict data policies across your entire team without touching any code. Gateway compliance features now include team-wide ZDR from the dashboard,

Use and manage Vercel Sandbox directly from the Vercel CLI

April 8, 2026

You can now manage Vercel Sandbox directly from the Vercel CLI using the vercel sandbox command. Create, manage and run sandboxes without leaving your terminal: Update the Vercel CLI to v54.15.1 or later and run vercel sandbox to get started. Learn more in the Vercel Sandbox CLI documentation. Read more

Summary of CVE-2026-23869

April 8, 2026

Summary A high-severity vulnerability (CVSS 7.5) in React Server Components can lead to Denial of Service. We created new rules to address these vulnerabilities and deployed them to the Vercel WAF to automatically protect all projects hosted on Vercel at no cost. However, do not rely on the WAF for full protection. Immediate upgrades to a patched version are required. Impact A specially crafted HTTP request can be sent to any App Router Server Function endpoint that, when deserialized, may trigger excessive CPU usage. This can result in denial of service in unpatched environments. These vulnerabilities are present in Next.js 13.x, 14.x, 15.x, 16.x and affected packages using the App Router. The issue is tracked upstream as CVE-2026-23869 Resolution After creating mitigations to address this vulnerability, we deployed them across our globally-distributed platform to protect our customers. We still recommend upgrading to the latest patched version. Updated releases of React and affected downstream frameworks include fixes to prevent this issue. All users should upgrade to a patched version as soon as possible. Fixed In = 15.0.0 to be fixed in 15.5.15 = 16.0.0 to be fixed in 16.2.3 Re

Chat SDK adds Liveblocks support

April 7, 2026

Chat SDK now supports Liveblocks, enabling bots to read and respond in Liveblocks Comments threads with the new Liveblocks adapter. This is an official vendor adapter built and maintained by the Liveblocks team. Teams can build bots that post, edit, and delete comments, react with emojis, and resolve @mentions within Liveblocks rooms. Try the Liveblocks adapter today: Read the documentation to get started, browse the directory, or build your own adapter. Read more

Opus 4.6 Fast Mode available on AI Gateway

April 7, 2026

Fast mode support for Claude Opus 4.6 is now available on AI Gateway. Fast mode is a premium high-speed option that delivers 2.5x faster output token speeds with the same model intelligence. This is an early, experimental feature. Fast mode's increased output token speeds enable new use cases, especially for human-in-the-loop workflows. Run large coding tasks without needing to context switch and get planning results without extended waits. To enable fast mode, pass speed: 'fast' in the anthropic provider options in AI SDK: You can use fast mode with Claude Code via AI Gateway by setting the CLAUDE_CODE_SKIP_FAST_MODE_ORG_CHECK variable in your shell configuration file or in ~/.claude/settings.json. Try fast mode directly in the AI Gateway playground for Opus 4.6. Fast mode is priced at 6x standard Opus rates. Standard Fast Mode Input: $5 / 1M tokens Output: $25 / 1M tokens Input: $30 / 1M tokens Output: $150 / 1M tokens All standard pricing multipliers (e.g., prompt caching) apply on top of these rates. Read more

GLM 5.1 on AI Gateway

April 7, 2026

GLM 5.1 from Z.ai is now available on Vercel AI Gateway. Designed for long-horizon autonomous tasks, GLM-5.1 can work continuously on a single task for extended periods, handling planning, execution, testing, and iterative refinement in a closed loop. Rather than one-shot code generation, it runs an autonomous cycle of benchmarking, identifying bottlenecks, and optimizing across many iterations, with particular strength in sustained multi-step engineering workflows. Beyond agentic coding, GLM-5.1 improves on general conversation, creative writing, front-end prototyping, and office productivity tasks like generating PowerPoint, Word, and Excel documents. To use GLM 5.1, set model to zai/glm-5.1 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Query and visualize workflow data in Vercel Observability

April 7, 2026

Observability Plus's query builder now lets you create custom queries on workflow runs and steps, visualizing traffic, performance, and other key metrics across Vercel Workflows. Queries include breakdowns by run and step status, and can be filtered and grouped by environment, project, workflow, and step. The query builder is available to Pro and Enterprise teams using Observability Plus. Learn more about Workflow SDK, Observability and Observability Plus. Read more

Manage Vercel Microfrontends with AI Agents and the CLI

April 7, 2026

Vercel Microfrontends now include two new setup and management tools: an AI skill for coding agents and new Vercel CLI commands. New Vercel Microfrontends skill: Install the Microfrontends skill to let your AI coding agent guide you through group creation with natural language prompts. It will automatically generate microfrontends.json, wire up framework integrations, and manage projects, all without leaving your editor. Once added, ask your agent to create your first microfrontend group using this prompt. Get started with the Microfrontends skill. New CLI commands: The Vercel CLI now includes commands for managing microfrontend groups, so you can create, inspect, and manage groups from the terminal without opening the dashboard. vercel microfrontends create-group vercel microfrontends inspect-group vercel microfrontends add-to-group vercel microfrontends remove-from-group vercel microfrontends delete-group Learn more in the CLI docs. Read more

Vercel CLI commands now scoped to local directory

April 6, 2026

Commands like vc project ls and vc domains ls now automatically use the scope of your linked local directory instead of defaulting to your global team. Previously, querying projects or domains inside a linked repository would return global results, creating an unexpected disconnect between your immediate working environment and the CLI output. This update aligns read-only commands with your local context, though you can still manually override the target team by passing the --scope flag. Run pnpm i -g vercel@latest to update to the latest Vercel CLI (at least v50.40.0). Read more

Query and manage Marketplace databases from the dashboard

April 6, 2026

You can now run SQL queries, view and edit data, and inspect your database schema directly from the Vercel dashboard. This capability is available for supported marketplace database integrations, including AWS Aurora Postgres, Neon, Prisma, and Supabase, with more services coming soon. Instead of relying on psql or external database UI tools, teams can manage their data directly within Vercel. The integration resource page for supported Postgres databases now includes three new tabs: Query: Run any SQL query, view results, and copy them as CSV, JSON, or Markdown. Data Editor: View table data in a spreadsheet-like interface. You can sort, copy, edit, insert, and delete rows, then apply changes to the database atomically. Schema: View tables and relations in a visual graph layout. This is currently available to users with Owner permissions. Manage your database by navigating to the Browser section of your database page in the dashboard. Read more

Team-wide Zero Data Retention and prompt training controls now on AI Gateway

April 6, 2026

AI Gateway now supports Zero Data Retention (ZDR) at the team level, removing the need to configure opt-outs or reach agreements with each provider individually. It routes requests only to providers where ZDR agreements are in place, with support for Anthropic, OpenAI, Google, and many more models. Team-wide ZDR Enable team-wide ZDR from the AI Gateway Dashboard Settings to enforce Zero Data Retention on every request your team makes, with no code changes required. Request-level ZDR & No Prompt Training You can also configure ZDR and/or No Prompt Training on the request level by setting zeroDataRetention: true or disallowPromptTraining: true per request. Enabling Zero Data Retention automatically includes training opt-out, so there's no need to configure both separately. Pricing Option Cost Availability Team-wide ZDR $0.10 per 1,000 requests Pro and Enterprise Per-request ZDR Free Pro and Enterprise Disallow Prompt Training (per-request) Free All users Each response includes metadata showing which providers were considered and which were filtered out, giving you an audit trail of policy enforcement. These controls work with the AI SDK, Chat Completions API, Responses API, Anthropic

Qwen 3.6 Plus on AI Gateway

April 2, 2026

Qwen 3.6 Plus from Alibaba is now available on Vercel AI Gateway. Compared to Qwen 3.5 Plus, this model adds stronger agentic coding capabilities, from frontend development to repository-level problem solving, along with improved multimodal perception and reasoning. It features a 1M context window and improved performance on tool-calling, long-horizon planning, and multilingual tasks. To use Qwen 3.6 Plus, set model to qwen/qwen3.6-plus in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Gemma 4 on AI Gateway

April 2, 2026

Gemma 4 26B (MoE) and 31B (Dense) from Google are now available on Vercel AI Gateway. Built on the same architecture as Gemini 3, both open models support function-calling, agentic workflows, structured JSON output, and system instructions. Both support up to 256K context, 140+ languages, and native vision. 26B (MoE): Activates only 3.8B of its 26B total parameters during inference, optimized for lower latency and faster tokens-per-second. 31B (Dense): All parameters are active during inference, targeting higher output quality. Better suited as a foundation for fine-tuning. To use Gemma 4, set model to google/gemma-4-31b-it or google/gemma-4-26b-a4b-it in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Zero-configuration Go backend support

April 2, 2026

Go API backends can now be deployed on Vercel with zero-configuration deployment. Vercel now recognizes Go servers as first-class backends and automatically provisions the right resources and configures your application without redirects in vercel.json or the /api folder convention. Backends on Vercel use Fluid compute with Active CPU pricing by default. Your Go API scales automatically with traffic, and you pay only for active CPU time rather than idle capacity. Read more

Optimizing Vercel Sandbox snapshots

April 2, 2026

When we recently shipped filesystem snapshots in Vercel Sandbox to let teams capture and restore a sandbox's entire filesystem state, our initial engineering focus was entirely on reliability, making sure the system would never fail to snapshot or lose data. Once that foundation was stable, our attention turned to performance. p75 snapshot restores were taking over 40 seconds, and through parallelization and local caching, we brought that under one second. What a snapshot looks like on disk Vercel Sandbox runs on the same infrastructure as our internal builds product, Hive. Each sandbox is an isolated container inside a Firecracker microVM. A snapshot is a compressed copy of the sandbox's disk. We're working with two different files: The raw disk image (.img), which can be several GBs A compressed version in our custom VHS format (Vercel Hive Snapshot), which is what gets uploaded to and downloaded from S3 When you call sandbox.snapshot(), we compress the .img into a .vhs and upload it to S3. When you call Sandbox.create() with a snapshot, we download the .vhs and decompress it back. Without compression, every snapshot operation transfers hundreds of MBs to low GBs over the network

Custom Class Serialization in Workflow SDK

April 2, 2026

Workflow SDK now supports custom class serialization, letting you pass your own class instances between workflow and step functions. Workflow SDK serializes standard JavaScript types like primitives, objects, arrays, Date, Map, Set, and more. Custom class instances were previously not supported because the serialization system didn't know how to reconstruct them. With the new @workflow/serde package, you can define how your classes are serialized and deserialized by implementing two static methods using WORKFLOW_SERIALIZE and WORKFLOW_DESERIALIZE. Here's an example of how we used custom serialization in @vercel/sandbox to greatly improve DX: Once implemented, instances of your class can be passed as arguments and return values between workflow and step functions, with the serialization system handling conversion automatically. See the full example application utilizing @vercel/sandbox and Workflow SDK. Read the serialization documentation to learn more. Read more

Build an MCP server with Nuxt

April 2, 2026

Developers building AI features with Nuxt can now create Model Context Protocol (MCP) servers directly within their applications using the Nuxt MCP Toolkit. Install the module The module lets you define tools with Zod validation, expose data as resources, and create reusable prompts. It also includes an integrated MCP Inspector for debugging, middleware support, dynamic definitions, session persistence across tool calls, and a Code Mode that lets models orchestrate multiple tool calls in a single execution. Read the documentation or the walkthrough guide to start building. Read more

Base fee for Observability Plus removed

April 2, 2026

We've eliminated the $10 base fee for Observability Plus for all customers. We now only charge you for the observability events collected. Observability Plus gives you full visibility into your Vercel data with enhanced logs and metrics, anomaly alerts, 30-day log retention, and the context to debug issues fast. Start using Observability Plus today. Read more

GLM 5V Turbo on AI Gateway

April 1, 2026

GLM 5V Turbo from Z.ai is now available on Vercel AI Gateway. GLM 5V Turbo is a multimodal coding model that turns screenshots and designs into code, debugs visually, and operates GUIs autonomously. It's strong at design-to-code generation, visual code generation, and navigating real GUI environments, at a smaller parameter size than comparable models. To use GLM 5V Turbo, set model to zai/glm-5v-turbo in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in custom reporting, observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

How Waldium made a blog platform work for humans and AI alike

April 1, 2026

Waldium is a two-person, YC-backed startup that built an agentic CMS for businesses. Co-founded by Amrutha Gujjar and CTO Shivam Singhal, the platform automates content research and creation, and gives every customer blog its own MCP server endpoint so AI agents can query it directly. Today, it all runs from a single Next.js deployment on Vercel. Amrutha spent years building software where the infrastructure kept getting in the way. "I was never stuck on the part I was excited to build," she says. "It was always the infrastructure friction around it." That frustration became a design principle for Waldium. The infrastructure needed to disappear so the product could stay focused on what made it interesting. Waldium on Vercel 500+ customer blogs served from a single Vercel deployment New customers get MCP endpoints live in under 5 minutes AI query response times consistently under 50ms globally 45% lower infrastructure cost vs. per-customer deployment model One customer generated 1,000+ posts in a single month; research time cut from weeks to hours When agents became part of the audience Waldium started the way most content platforms do, building blogs for humans to read. But somethi

Chat SDK adds Zernio support

April 1, 2026

Chat SDK now supports Zernio, a unified social media API, with the new Zernio adapter. This is an official vendor adapter built and maintained by the Zernio team. Teams can build bots that work across Instagram, Facebook, Telegram, WhatsApp, X/Twitter, Bluesky, and Reddit through a single integration. Try the Zernio adapter today: Feature support varies by platform; rich cards work on Facebook, Instagram, Telegram, and WhatsApp, while editing and streaming are currently limited to Telegram. Read the documentation to get started, browse the directory, or build your own adapter. Read more

Vercel Queues now supports 7-day message TTL

April 1, 2026

Vercel Queues now supports a maximum message TTL and delivery delay of 7 days, up from 24 hours. The default message TTL of 24 hours and delivery delay of 0 seconds remain unchanged. These expanded limits enable longer retention windows and support background work patterns for delayed tasks, scheduled workflows, and long-running jobs. Send message example Explore the documentation. Read more

Axios package compromise and remediation steps

March 31, 2026

The axios npm package was compromised in an active supply chain attack discovered on March 31, 2026. Vercel investigated this issue and implemented remediation actions to protect the platform. No Vercel systems were affected. The npm registry removed the compromised package versions, and the latest tag now points to the safe axios@1.14.0 release. We’ve blocked outgoing access from our build infrastructure to the Command & Control hostname sfrclak.com. The malicious version of the package has been blocked and unpublished from npm. Vercel’s own infrastructure and applications have been unaffected. We recommend checking your supply chain for exposure. Affected versions Projects using axios@1.14.1 or axios@0.30.4 in their build environments are affected by this vulnerability. Check your dependencies and lockfiles for: axios@1.14.1 axios@0.30.4 plain-crypto-js@4.2.1 Resolution If your deployments used the malicious package version listed above in your build environment, take the following actions: Search your lockfiles and node_modules for plain-crypto-js to identify compromised installations Redeploy your project to ensure your build uses a clean version of axios Rotate API keys, datab

How FLORA shipped a creative agent on Vercel's AI stack

March 31, 2026

FLORA on Vercel 2x faster to production with their generation system Zero infrastructure debates after migration 50+ image models orchestrated A seasonal fashion launch is a story, not a single frame. Crafting that story is a process of exploration: It’s the same piece, worn by different models. The same pose, with different lighting and angles. The same set, with a background that shifts from glossy to gritty. FLORA was built to make that kind of visual iteration available to anyone through a digital canvas. Their new creative agent FAUNA acts like a design partner, turning ideas into a map of creative directions. From FLORA to FAUNA: a creative agent Flora started as a node-based creative workflow canvas. For advanced users, that canvas is powerful: you can create steps and branches, adding detailed prompts at each stage to hone every detail of the images. But the canvas also came with a tradeoff: it asked every creative to think like a workflow designer. You had to configure each state of an exploration with prompts and model choices, interrupting the creative process. That’s why they built FAUNA. The agent removes the setup burden, but keeps the full power of AI. Instead of sta

Transfer Marketplace resources between teams

March 31, 2026

You can now transfer Marketplace resources between teams directly from the Vercel dashboard without relying on the API. This simplifies resource management during team or project changes. Both owner and member roles on the source and destination teams can initiate transfers. The destination team must have the corresponding integration installed before receiving a resource. The feature currently supports transfer databases from Prisma, Neon and Supabase , with additional providers and product support coming soon. Start from your database settings in the dashboard, or learn more in the documentation. Read more

Vercel CDN now respects Cache-Control headers from external origins by default

March 30, 2026

New Vercel projects will honor cache-control headers by default when proxying requests to external origins, starting April 6th. Previously, responses served through rewrites to external origins were uncached by default, and enabling caching required the x-vercel-enable-rewrite-caching header in vercel.json. Now, Vercel's CDN automatically respects your origin's caching headers. What's changing: For new projects, Vercel will cache responses from external origins according to upstream Cache-Control, CDN-Cache-Control and Vercel-CDN-Cache-Control headers by default. You can use Cache Tags (Vercel-Cache-Tag header) from your origins to purge cached content. Existing projects can opt in to the new caching behavior from the project dashboard. You can opt out of caching for specific request paths by setting the x-vercel-enable-rewrite-caching header to 0. Review your upstream cache headers before April 6th when creating a new project that proxies to external origins without caching, ensuring they reflect your intended caching strategy. Learn more about Rewrites to External Origins and configure routing in the CDN tab of your project settings. Read more

Agent responsibly

March 30, 2026

The following is based on an internal talk given at Vercel. We're sharing it publicly because the problem it describes isn't unique to us, and the framework is useful for any team shipping with agents. Coding agents generate code at unprecedented speeds. In the hands of disciplined engineers, they are a productivity multiplier. But without rigorous judgment, they are a highly efficient way to ship bad assumptions directly to production. When teams deploy agent-generated code blindly, the fallout can be immediate and severe. A flawless-looking pull request can ship a query that passes tests, but scans every row in production. Retry logic that seems correct can cause a thundering herd on a downstream service. And a cache with no TTL can quietly grow until Redis dies. Green CI is no longer proof of safety. In an agentic world, passing CI is merely a reflection of the agent's ability to persuade your pipeline that a change is safe, even if it will immediately degrade your infrastructure at scale. False confidence Agent-generated code is dangerously convincing. It comes with a polished PR description, passes static analysis, follows repository conventions, and includes reasonable test c

Making Turborepo 96% faster with agents, sandboxes, and humans

March 30, 2026

Turborepo is now 81-91% faster to compute its task graph in our repositories, scaling with repo size. On our 1,000+ package monorepo, turbo run now feels instant. Time to First Task is now 11x faster. After testing my changes with some open source Turborepos and asking Vercel customers to try canary releases on their repositories, I found the performance improvement could get as high as 96% depending on the size and complexity of the repository. The process behind earning these performance gains is worth sharing, because it wasn't one optimization or one technique. It was eight days of mixing AI agents, Vercel Sandboxes, and typical, boring engineering practices. How Turborepo schedules your tasks Every turbo run starts by analyzing your monorepo's structure, scripts, and dependencies to build a task graph. That graph determines execution order, creates parallelism, and powers caching so you never repeat the same work twice. Building the task graph is overhead you pay before your repository's work begins. The larger the repo, the higher the cost. On our 1,000-package monorepo, that cost was around 10 seconds on an M4 Pro Max. I don't know about you, but I found that unacceptable. S

Vercel plugin now supported on OpenAI Codex and Codex CLI

March 26, 2026

The Vercel plugin now supports OpenAI Codex and the Codex CLI. With the plugin, teams can access over 39 platform skills, three specialist agents, and real-time code validation to assist with their development workflow. Install it in the Codex app or from the Codex CLI: Learn more about the plugin in the documentation. Read more

Automatic persistence now in beta on Vercel Sandbox

March 26, 2026

Vercel Sandboxes can now automatically save their filesystem state when stopped and restore it when resumed. This removes the need for manual snapshots, making it easier to run long-running, durable sandboxes that continue where you left off. How it works A sandbox is the durable identity, now identified by a name, its filesystem state, and configuration options. A session is the compute tied to that state, invoked as needed. Automatic persistence introduces orchestration that separates storage from compute, reducing the need for manual snapshotting, so: when you stop a sandbox, the session shuts down but the filesystem is automatically snapshotted. when you resume, a new session boots from that snapshot. This state storage is not charged, so you pay when your setting is active. Persistence is enabled by default in the beta SDK, can be disabled between sessions with persistent: false. When disabled, the sandbox still exists after being stopped and can be resumed by its name, but each session starts with a clean filesystem. If a sandbox is stopped and you run a command, the SDK will transparently create a new session, so you don't need to check state or manually restart The beta SDK

Vercel Sandboxes now allow unique, customizable names

March 26, 2026

Vercel Sandboxes created with the latest beta package will now have a unique, customizable name within your project, replacing the previous ID-based identification. Names make sandboxes easy to find and reference: The beta CLI adds configuration management and session inspection: Named sandboxes are the mechanism for identifying automatic persistence, which allows your session to be more easily identified for at both time of creation resumption. Install the beta packages to try named sandboxes today: pnpm install @vercel/sandbox@beta for the SDK, and pnpm install -g sandbox@beta for the CLI. Learn more in the documentation. Read more

Unified reporting for all AI Gateway usage

March 25, 2026

If you're shipping AI features, you already have usage data. The problem is that it's split across providers, keys, and dashboards, so it's hard to answer basic questions before the bill shows up. You've probably felt the drift into after-the-fact reconciliation. Provider consoles only show their own slice, so you end up exporting CSVs, rebuilding views in spreadsheets, and still missing the context that matters, like your tags, feature boundaries, and internal user IDs. When BYOK enters the picture, it gets worse because spend and usage scatter across whatever keys your users bring. The Custom Reporting API for AI Gateway is now available in beta for teams on the Pro and Enterprise plans. It gives you programmatic access to cost, token usage, and request volume across your AI Gateway traffic, including BYOK requests. You can break down spend by model, provider, user ID, custom tag, or credential type. That makes it possible to track costs and usage per feature, per end customer, and per pricing tier from a single endpoint. You can also query it live via Claude Code. How a platform saved $80K One AI platform aggregating models for 200K+ users previously relied on a separate proxy l

Elastic build machines now available in beta

March 24, 2026

Elastic build machines are now available in beta for all paid plans, giving teams control over build performance without project-level micromanagement. You can configure elastic builds at the team or project level. Rather than a one-size-fits-all approach, Vercel evaluates each project individually and assigns the right machine for its actual needs. Smaller, simpler projects may benefit from cost-efficient Standard build machines while more complex workloads can automatically scale up to Enhanced or Turbo machines. This smart assignment prevents over-provisioning, with teams automatically getting optimal performance at the right cost for every project. Enable elastic builds in your team settings or project settings, or read the builds documentation. Read more

Chat SDK now supports concurrent message handling

March 24, 2026

Chat SDK now lets you control what happens when a new message arrives before a previous one finishes processing, with the new concurrency option for the Chat class. Four strategies are available: drop (default): discards incoming messages queue: processes the latest message after the handler finishes debounce: waits for a pause in conversation, processes only the final message concurrent: processes every message immediately, no locking Read the documentation to get started. Read more

Chat SDK now supports scheduled Slack messages

March 24, 2026

Chat SDK now supports scheduled messages on Slack, allowing you to deliver a message at a future time. Use thread.schedule() and pass your message and a postAt date, like: Read the documentation to get started, and learn more about the Slack adapter. Read more

new.website joins forces with v0

March 23, 2026

v0 and new.website have joined forces to accelerate our vision of helping anyone ship complete, production-ready software with AI. new.website was founded to make it effortless to create beautiful websites with all the tools included, from built-in forms to SEO. They’re joining the v0 team to help shape our built-in tools and integrations. Why new.website The team at new.website built thoughtful primitives that exposed configuration for critical website features instead of requiring users to implement them with multiple prompts and code checks. Here are examples of their built-in tooling: Built-in forms and databases. Every site came with forms by default and every submission was automatically stored without any setup. Simple SEO features. Website generation included core SEO configuration like meta tags and structure, and those settings were exposed in the UI for direct configuration. Content management. Content could be edited outside the website, even when generated by an agent, which was a better user experience for long-form entries like blogs. The goal of every primitive was to reduce the amount of prompting required to implement and configure baseline website functionality.

SERHANT.'s playbook for rapid AI iteration

March 23, 2026

Impact at a glance Started with Next.js on Vercel, which made it easier to expand to a React Native iOS app without rebuilding their backend Engineers focus on AI design and iteration instead of platform plumbing Orchestrates OpenAI, Claude, and Gemini by task to optimize cost vs output Scaled from an internal pilot to 800–900+ real estate agents without replatforming When Jeremy Bunting joined SERHANT. as VP of Engineering in February 2024, S.MPLE was already showing promise. 200 real estate agents were piloting the AI product, which was designed to save time by automating cumbersome and repetitive daily tasks, like market analysis and contact management. S.MPLE was a Next.js progressive web app deployed on Vercel, and that foundation gave the team leverage. They could keep the API layer steady while expanding the client experience, including expansion to a React Native iOS app, all without a backend rebuild. But Bunting had a problem that keeps many engineering leaders up at night: the AI landscape changes faster than most teams can implement infrastructure updates. The team needed to move fast, scale confidently, and stay flexible enough to swap models, add new capabilities, and

Enterprise teams can now set their default build machine

March 23, 2026

Enterprise team owners can now set a default build machine at the team level. This setting automatically applies to newly created projects, though you can still override it on a per-project basis. Existing projects retain their current configurations unless you explicitly choose to apply the new team default to all of them when saving. Learn more about build machines or try it out from settings. Read more

Activity Log now available in Vercel CLI

March 23, 2026

The activity log provides a list of all activities on a team, along with the user who performed the event, the type of event, and time. We've now added the vercel activity command to the CLI so that you can query for activity events. You can filter events by type, date range, and project, or view all events across your team. Learn more about the Vercel CLI and the vercel activity command. Read more

View specific error codes in runtime logs

March 23, 2026

You can now view specific error codes in runtime logs. Runtime logs helps you view and troubleshoot errors in your applications on Vercel. In addition to the HTTP status response code, we also list the specific error code the request details panel of the runtime logs page of the Vercel dashboard. This makes it easier to diagnose why a request failed. Learn more about errors on Vercel. Read more

Sandbox SDK adds file permission control

March 20, 2026

Vercel Sandbox SDK 1.9.0 now supports setting file permissions directly when writing files. By passing a mode property to the writeFiles API, you can define permissions in a single operation. This eliminates the need for an additional chmod execution round-trip when creating executable scripts or managing access rights inside the sandbox. See the documentation to learn more. Read more

Build knowledge agents without embeddings

March 19, 2026

Most knowledge agents start the same way. You pick a vector database, then build a chunking pipeline. You choose an embedding model, then tune retrieval parameters. Weeks later, your agent answers a question incorrectly, and you have no idea which chunk it retrieved or why that chunk scored highest. We kept seeing this pattern internally and for teams building agents on Vercel. The embedding stack works for semantic similarity, but it falls short when you need a specific value from structured data. The failure mode is silent: the agent confidently returns the wrong chunk, and you can't trace the path from question to answer. That's why we tried something different. We replaced our vector pipeline with a filesystem and gave the agent bash. Our sales call summarization agent went from ~$1.00 to ~$0.25 per call, and the output quality improved. The agent was doing what it already knew how to do: read files, run grep, and navigate directories. So we open-sourced the Knowledge Agent Template, a production-ready version of this architecture built on Vercel. What the template does The Knowledge Agent Template is an open source, file-system-based agent you can fork, customize, and deploy.

Two startups at global scale without DevOps

March 19, 2026

Leonardo.AI processes more than 4.5 million images every day across cities worldwide, and Relevance AI's agents run autonomously across time zones, touching Salesforce, HubSpot, Slack, and dozens of other systems without pause. Neither company has a dedicated DevOps team. That's not an oversight. It's an operational model. The startup ecosystem in APAC is a clear example of why this model is taking hold. AI-native startups are surging across the region: over 1,000 are building in Australia alone, according to the State of Australian Startup Funding Report. Meanwhile, nearly a third of Singapore's new startups last year were focused on AI. Startups can find capital, but not DevOps talent Billions in AI venture capital are flowing into APAC. In Australia alone, more than $1 billion was invested in AI-native startups in 2025. Singapore attracted $8.4 billion, accounting for 75% of the region's total. The money is there. The infrastructure talent is not. A single DevOps engineer in Australia costs $150K+, takes months to close, and that's only if you can find one. Sixty to 80 percent of APAC organizations report difficulty filling IT roles, according to IDC. For a 15-person startup, th

Vercel is now available in Stripe Projects

March 19, 2026

You can now signup and deploy to Vercel through Stripe Projects. Available in developer preview, this CLI-based workflow lets teams and AI agents create infrastructure environments directly from the terminal. As a launch and design partner for Stripe Projects, Vercel enables a unified provisioning flow. Using the CLI, you can: Connect existing Vercel account or a new will be created for you Provision new Vercel projects directly from the terminal Deploy and host applications on Vercel as part of a unified infrastructure setup This integration reduces friction between local development and cloud hosting, allowing infrastructure to be provisioned programmatically. To start provisioning your infrastructure programmatically, install the Stripe CLI plugin and initialize your project: Select Vercel to deploy and start building, or learn more about Stripe Projects. Read more

Chat SDK brings agents to your users

March 19, 2026

In early March, we gave the entire organization a challenge: figure out how to multiply your output. People created agents. Mostly chat bots, but dedicated ones, purpose-built for real workflow augmentation: the agents were doing things automatically that would otherwise be tedious and time consuming. Initially people built individual interfaces for their agents, and AI SDK made that easy with out-of-the box model integrations and AI Elements to simplify UI design. Then we hit a constraint. People wanted to interact with the agents in Slack, meaning everyone had to figure out how to integrate with Slack's API. Then the problem got worse. Once the agents were in Slack, people wanted to integrate them with even more platforms, like Discord, Github, and even issue trackers like Linear. Each of those introduced a new integration adventure for every agent. It hit us. Instead of asking people to come to agents, we needed to deliver agents to the places they were already working. Chat needs integration abstraction We realized we had made it easy for our teams to create agents, but we needed to make it easier to extend them for adoption across platforms. This is true for every company. Peo

v0 now includes diff view to review code changes

March 18, 2026

v0 now includes a dedicated diff view to review code changes directly within the interface. You can see exactly what was modified file by file, complete with line addition and deletion counts. Check out v0.app today or read the documentation to learn more. Read more

MiniMax M2.7 is live on AI Gateway

March 18, 2026

MiniMax M2.7 is now available on Vercel AI Gateway in two variants: standard and high-speed. M2.7 is a major step up from previous M2-series models in software engineering, agentic workflows, and professional office tasks. The model natively supports multi-agent collaboration, complex skill orchestration, and dynamic tool search for building agentic workflows. M2.7 also improves on production debugging and end-to-end project delivery. The high-speed variant delivers the same performance for 2x the cost of standard at ~100 tokens per second for latency-sensitive use cases. To use M2.7, set model to minimax/minimax-m2.7 or minimax/minimax-m2.7-highspeed in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

360 billion tokens, 3 million customers, 6 engineers

March 18, 2026

Impact at a glance Durable ships new production agents to customers in a single day AI features and agents serve ~1.1B tokens per day (360B per year) 10x leverage for every engineer, product manager, and designer 3-4x lower infra cost compared to self hosting Durable began with a simple goal: make owning a business easier than having a job. 60% of U.S. adults say they want to be their own boss, but only about 4% actually do it. Durable’s bet is that the blocker isn’t ambition. It’s friction. “Small businesses are death by a thousand tools, logins, workflows, and designs,” explained James Clift, founder of Durable. “If you remove those barriers, business owners can focus on their customers.” Today, Durable is an AI business builder that helps entrepreneurs launch in minutes, then optimize with agents that handle things like SEO, content, and operations. The gap between idea and ownership has never been smaller. It feels like one seamless experience to their customers, but under the hood it’s a multi-tenant, multi-product platform that has to run millions of individual businesses safely, reliably, and cost-effectively. As they scaled, manually operating multiple services just to self

End-to-end encryption for Vercel Workflow

March 17, 2026

Vercel Workflow now encrypts all user data end-to-end without requiring any code changes. Workflow inputs, step arguments, return values, hook payloads, and stream data are automatically encrypted before being written to the event log. This makes it safe to pass sensitive data, such as API keys, tokens, or user credentials, across boundaries. The event log only ever stores ciphertext, while your step functions work exactly as before. Your workflow and step functions work exactly as before; all data flowing through the event log is encrypted automatically. Each Vercel deployment receives a unique encryption key. The key derivation and encryption stack works as follows: Each workflow run derives its own key via HKDF-SHA256 Data is encrypted with AES-256-GCM to ensure confidentiality and integrity Encrypted fields display as locked placeholders in the dashboard until decrypted You can access encrypted data through two methods: Web dashboard: Click the Decrypt button in the run detail panel. Decryption happens entirely in the browser via the Web Crypto API, so the observability server never sees your plaintext data. CLI: Add the --decrypt flag to the inspect command. Decryption follows

Vercel Open Source Program: Winter 2026 cohort

March 17, 2026

In April, we launched the Vercel Open Source Program to give maintainers the resources, credits, and support they need to ship faster and scale confidently. The first group joined through our spring 2025 cohort. Today we are welcoming the winter 2026 cohort. From AI-native apps and developer infrastructure to design systems and creative tooling, open source builders continue to amaze us. Meet the creators and explore their projects. Winter 2026 Cohort Answer Overflow Answer Overflow is an open source project that makes the content sent in public Discord servers accessible to search engines and AI agents. Over 300+ communities use Answer Overflow to create web pages from their Discord content and it serves 1.5 million monthly active users. Creator: Rhys Sullivan View the project View GitHub Repository UI TripleD UI TripleD offers UI blocks, components, and full pages in shadcn/ui and Base UI, powered by Framer Motion, with a Landing Builder, Background Builder, and Grid Generator. Creator: Moumen Soliman View the project View GitHub repository hot-updater An open source React Native OTA (Over-The-Air) update solution built as a modern alternative to CodePush, designed to avoid vendo

Introducing the Vercel plugin for coding agents

March 17, 2026

Claude Code and Cursor can now further understand Vercel projects using the new Vercel plugin and a full platform knowledge graph. The plugin observes real-time activity, including file edits and terminal commands, to dynamically inject Vercel knowledge into the agent's context. Key capabilities include: Platform knowledge: Access 47+ skills covering the Vercel platform, including Next.js, AI SDK, Turborepo, Vercel Functions, and Routing Middleware, powered by a relational knowledge graph Specialized tooling: Use three specialist agents (AI Architect, Deployment Expert, Performance Optimizer) and five slash commands (/bootstrap, /deploy, /env, /status, /marketplace) Context management: An injection engine and project profiler rank, deduplicate, and budget-control loaded context Code validation: PostToolUse validation catches deprecated patterns, sunset packages, and stale APIs in real time Instead of standard retrieval, the plugin compiles pattern matchers at build time and runs a priority-ranked injection pipeline across seven lifecycle hooks. Skills fire when glob patterns, bash regexes, import statements, or prompt signals match, and are then deduplicated across the session to e

Updates to Terms of Service

March 17, 2026

Agents are reshaping the tools developers use, the applications they build, and the infrastructure that runs them. We’ve updated our Terms of Service and Privacy Policy to reflect how Vercel uses data to support agentic features, improve our platform, and contribute to the AI ecosystem. What is changing? Agentic infrastructure capabilities We are developing features that allow Vercel to do more to keep your apps running efficiently, including: Proactively investigating and mitigating incidents Analyzing web app performance data and suggesting improvements Identifying where your spend is going and creating PRs to optimize usage Vercel may also use data to help improve our tools to fight fraud and abuse of the Vercel platform. Optional AI model training You may choose whether to allow Vercel to: Use your code and Vercel agent chats to improve Vercel models Share your code and Vercel agent chats with AI model providers Defaults by plan for optional AI model training: Hobby (including Trial Pro): Opted in for AI model training by default, with self-serve opt-out in Team and Project Settings Pro (paid): Opted out of AI model training by default, with self-serve opt-in in Team and Projec

Use GPT 5.4 Mini and Nano on AI Gateway

March 17, 2026

GPT-5.4 Mini and GPT-5.4 Nano from OpenAI are now available on Vercel AI Gateway. Both models deliver state-of-the-art performance for their size class in coding and computer use, and are built for sub-agent workflows where multiple smaller models coordinate on parts of a larger task. The models also support the verbosity and reasoning level parameters, giving you control over response detail and how much the model reasons before answering. GPT-5.4 Mini GPT-5.4 Mini handles code generation, tool orchestration, and multi-step browser interactions more reliably than previous mini-tier models. It's a strong default for agentic tasks that need to balance capability and cost. To use this model, set model to openai/gpt-5.4-mini in the AI SDK. GPT-5.4 Nano GPT-5.4 Nano performs close to GPT-5.4 Mini in evaluations at a lower price point. The model is well-suited for high-volume use cases like sub-agent workflows where cost scales with the number of parallel calls. To use this model, set model to openai/gpt-5.4-nano in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-p

New GitHub App permissions for Actions and Workflows

March 17, 2026

The Vercel GitHub App now requests two additional repository permissions on install: Actions (read) and Workflows (read & write). These permissions enable Vercel Agent to read workflow run logs to help diagnose CI failures and configure CI workflow files on your behalf. This also allows v0 to create complete, production-ready repositories with configured CI/CD pipelines. To use these features, accept the updated permissions in your GitHub organization or account settings. For full details on all permissions requested by the Vercel GitHub App, see the documentation. Read more

Streamdown 2.5 is now available

March 16, 2026

Streamdown is a React component library that makes rendering streaming markdown content easy and beautiful. Built for AI-powered applications, it handles the unique challenges that arise when markdown is tokenized and streamed in real time. v2.5 adds inline KaTeX support, staggered streaming animations, and a round of fixes for code blocks, CSV exports, and better Tailwind v3 compatibility. Streaming parser improvements The new inlineKatex option auto-completes $formula to $formula$ during streaming, avoiding ambiguity with currency symbols, and the option defaults to false. Block KaTeX completion is also fixed when streaming produces a partial closing $. Separately, single ~ between word characters (e.g. 20~25°C) is now escaped to prevent false strikethrough rendering, controlled via a new singleTilde option that is enabled by default. Staggered streaming animations Streaming word and character animations now cascade sequentially rather than animating all at once. The timing is configurable via a new stagger option (default 40ms). Set stagger: 0 to restore the previous behavior. Code blocks Custom renderers now receive the raw metastring from the code fence via a new optional meta

LiteLLM server now supported on Vercel

March 16, 2026

You can now deploy LiteLLM server on Vercel, giving developers LLM access with an OpenAI-compatible gateway connecting to any supported provider, including Vercel AI Gateway. To route a single model through Vercel AI Gateway, use the below configuration in litellm_config.yaml: Deploy LiteLLM on Vercel or learn more on our documentation Read more

Meet the 2026 Vercel AI Accelerator Cohort

March 16, 2026

The Vercel AI Accelerator is back, and this year we selected 39 early-stage teams from across the US, Europe, Asia, and Latin America to build with us for six weeks. The next generation of AI startups is building on our self-driving infrastructure, and the accelerator is how we work directly with the earliest-stage founders among them. This year’s cohort spans every industry, at varying points in their journey, but they share a clear point of view on what needs to exist right now and the urgency to ship it. Teams in the program get access to Vercel infrastructure, $8M+ in partner credits (AWS, Anthropic, OpenAI, Cartesia, ElevenLabs, and more), weekly sessions with founders and technical leaders, and dedicated VC mentors. Meet the 2026 cohort Alloy — Founder: Joe Harris Alloy Robotics turns terabytes of robot data into instant debugging, analysis, and reporting. Archival — Founders: Pablo Diaz-Granados, Jay Katyan, Griffin Ogura Archival reveals how teams are actually using AI. With workflow-level visibility and analytics, leaders can improve AI adoption and ROI. Ares — Founder: Tony Hunter Ares is a revenue command center, a single surface to find, engage, close, and grow revenue,

next-forge 6 is now available

March 13, 2026

next-forge is a production-grade Turborepo template for Next.js apps, designed to be a comprehensive, opinionated starting point for new apps. This major release comes with a number of DX improvements, an agent skill, and new guides for quickstart, Docker, and migration paths. next-forge skill You can now install a next-forge skill into your preferred agent, giving it structured knowledge of next-forge architecture, packages, and common tasks. Bun by default The default package manager is now Bun. The CLI init script detects your current package manager before prompting, and pnpm, npm, and yarn are still supported through the init flow. Graceful degradation Every optional integration now silently degrades when its environment variables are missing, rather than throwing an error. Stripe, PostHog, BaseHub, and feature flags all return safe defaults. The only required environment variable to boot the project is DATABASE_URL. New guides The quickstart guide gets you to a running dev server in a few minutes with just Clerk and a Postgres database. There is also a new Docker deployment guide and migration guides are available for Appwrite (auth, database, storage), Convex (database), and

How Notion Workers run untrusted code at scale with Vercel Sandbox

March 12, 2026

Notion Workers let you write and deploy code to give Custom Agents new powers: sync external data, trigger automations, call any API. With Workers, developers can build agents that sync CRM data on a schedule, open issues when error rates spike, and turn Slack threads into formatted content. Under the hood, every Worker runs on Vercel Sandbox. The problem: safely running code from any developer or agent Notion wanted to let anyone extend their platform with custom code. That's a hard infrastructure problem, but an even bigger security problem. Every Notion Worker runs arbitrary code generated by a third-party developer or agent, on behalf of a Notion user, potentially inside an enterprise workspace. Without proper isolation, a Worker would run in the same environment as the Custom Agent, with access to its secrets, permissions, and everything else in that execution context. A single prompt injection could exfiltrate credentials or access another user's data. The requirements were clear: Hard isolation: One Notion Worker can never access another's data or state Credential security: Notion Workers need API keys to talk to external services, but those secrets can never be exposed to t

Vercel now supports Domain Connect as a DNS Provider

March 12, 2026

Vercel now supports Domain Connect as a DNS provider, enabling external services to configure Vercel domains. Teams that use Vercel as their DNS host can set up their domain in one click without manually copying DNS records. This provides faster setup, fewer copy-and-paste mistakes, and less provider-specific documentation. We are launching this capability with Resend. When configuring email for a custom domain in Resend, teams can automatically provision the necessary DNS records directly on their associated Vercel domain. This update streamlines domain management for domains purchased or transferred to Vercel. To request support for a specific Domain Connect template, contact our team. To learn more about Domain Connect, read the docs. Read more

AI Elements 1.9 is now available

March 12, 2026

AI Elements 1.9 adds new components, an agent skill, and a round of bug fixes across the library. AI Elements skill You can now install an AI Elements skill into your preferred agent, giving it a better understanding of how to build and use composable AI interfaces. The new component renders JSX strings dynamically, supporting streaming scenarios where JSX may be incomplete. It automatically closes unclosed tags during streaming, making it a good fit for displaying AI-generated UI in real time. A new sub-component that captures a screenshot of the current page, useful for giving visual feedback to AI models. Download conversations The component now includes an optional button that downloads the conversation as a markdown file. Read the documentation to get started. Read more

Deprecating the DHE cipher suite for TLS connections

March 12, 2026

On June 30th, 2026, Vercel will remove support for the legacy DHE-RSA-AES256-GCM-SHA384 cipher suite. This cipher may still be used by automated systems, security scanners, and HTTP clients with non-standard TLS configurations. After this date, clients using TLS 1.2 will only be able to connect to the Vercel network with our six remaining cipher suites: ECDHE-ECDSA-AES128-GCM-SHA256 ECDHE-RSA-AES128-GCM-SHA256 ECDHE-ECDSA-AES256-GCM-SHA384 ECDHE-RSA-AES256-GCM-SHA384 ECDHE-ECDSA-CHACHA20-POLY1305 ECDHE-RSA-CHACHA20-POLY1305 Modern clients and TLS 1.3 connections are unaffected. If you operate integrations or automated systems that connect to a domain hosted on Vercel over TLS 1.2, verify that your TLS client supports at least one of the above cipher suites. Modern TLS libraries support these by default. Read more

Vercel Flags are now optimized for agents

March 11, 2026

The Vercel CLI now supports programmatic flag management, giving teams a direct way to create and manage feature flags from the terminal without opening the dashboard. Add the Flags SDK skill Building on this foundation, the Flags SDK skill lets AI agents generate and manage flags through natural language prompts. The skill leverages the CLI under the hood, enabling agents to implement server-side evaluation that prevents layout shifts and maintains confidentiality. Using the SDK's adapter pattern, agents can connect multiple providers and evaluate user segments without rewriting core flag logic. Once added, try prompting your agent with this prompt to create your first flag. Start generating flags with the Flags SDK skill. Read more

Subscribe to webhook events for Vercel Flags

March 11, 2026

You can now subscribe to webhook events for deeper visibility into feature flag operations on Vercel. New event categories include: Flag management: Track when teams create, modify, or delete flags across your project. Segment management: Receive alerts when segments are created, updated, or deleted. These events help teams build monitoring directly into their workflows. You can track the complete lifecycle of your flags, monitor changes across projects, and integrate feature flag data with your external systems. Read the documentation to start tracking feature flag events. Read more

Chat SDK adds WhatsApp adapter support

March 11, 2026

Chat SDK now supports WhatsApp, extending its single-codebase approach to Slack, Discord, GitHub, Teams, and Telegram with the new WhatsApp adapter. Teams can build bots that support messages, reactions, auto-chunking, and read receipts. The adapter handles multi-media downloads (e.g., images, voice messages, stickers) and supports location sharing with Google Maps URLs. Try the WhatsApp adapter today: The adapter does not support message history, editing, or deletion. Cards render as interactive reply buttons with up to three options, and fall back to formatted text. Additionally, WhatsApp enforces a 24-hour messaging window, so bots can only respond within that period. Read the documentation to get started, browse the directory, or build your own adapter. Special thanks to @ghellach, whose community contribution in PR #102 laid the groundwork for this adapter. Read more

Improved data collection for Web Analytics and Speed Insights with resilient intake

March 11, 2026

Web Analytics and Speed Insights version 2 introduces resilient intake to improve data collection reliability. By dynamically discovering endpoints instead of relying on a single predictable path, the new packages ensure you capture more complete traffic and performance data. To utilize resilient intake, update your packages and deploy your changes. No other configuration is required, and existing implementations will continue working as before. It's available to all teams at no additional cost. Install the latest versions These packages include a license change from Apache-2.0 to MIT to align with other open source packages. Nuxt applications can leverage Nuxt modules for a one-line installation of Speed Insights and Web Analytics. Update your packages to capture more data, or explore the Web Analytics documentation and Speed Insights documentation. Read more

Try Grok 4.20 on AI Gateway

March 11, 2026

Grok 4.20 is now available on Vercel AI Gateway in three variants: Reasoning, Non-Reasoning, and Multi-Agent. As xAI's newest flagship model, Grok 4.20 delivers industry-leading speed and agentic tool calling capabilities. It offers a low hallucination rate and strict prompt adherence, producing precise responses. Grok 4.20 Non-Reasoning: Non-reasoning variant optimized for speed and direct responses Grok 4.20 Reasoning: Extended thinking for complex problem-solving tasks Grok 4.20 Multi-Agent: Purpose-built for multi-agent orchestration and collaboration To use Grok 4.20, set model to xai/grok-4.20-non-reasoning-beta, xai/grok-4.20-reasoning-beta or xai/grok-4.20-multi-agent-beta in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Vercel Sandbox now supports 1 vCPU + 2 GB RAM configurations

March 10, 2026

Vercel Sandbox now supports creating Sandboxes with only 1 vCPU and 2 GB of RAM. This is ideal for single-threaded or light workloads which don't benefit from additional system resources. When unspecified, the default is still 2 vCPUs and 4 GB of RAM. Get started by setting the resources.vcpus option in the SDK: Or using the --vcpus option in the CLI: Learn more about Sandbox in the docs. Read more

Chat SDK now has an adapter directory

March 10, 2026

Chat SDK now has an adapter directory, so you can search platform and state adapters from Vercel and the community. These include: Official adapters: maintained by the core Chat SDK team and published under @chat-adapter/* Vendor-official adapters: built and maintained by the platform companies, like Resend and Beeper. These live in their GitHub org and are documented in their docs. Community adapters are built by third-party developers, and can be published by one, following the same model as AI SDK community providers. We encourage teams to build and submit adapters to be included in this new directory, like Resend's adapter that connects email to Chat SDK: Browse the adapter directory or read the contributing guide to learn how to build, test, document, and publish your own adapter. Read more

How we run Vercel's CDN in front of Discourse

March 10, 2026

Vercel's CDN can front any application, not just those deployed natively on the platform, and it's simple to set up. This allows you to add firewall protection, DDoS mitigation, and observability to platforms like Discourse or WordPress without migrating them completely. The Vercel Community is an example of this architecture. It is a Discourse application hosted elsewhere, but we proxy it ourselves via Vercel's CDN, which both protects the app and gives us access to useful features in Vercel's website stack: Web Analytics gives us anonymized, cookie-free demographic and referrer data, so we can see where users are coming from and what they're looking for. Firewall gives us DDoS protection and has automatically prevented several attacks in the last year. Bot Management lets us block malicious scrapers while allowing trusted crawlers to index the forum and allow community posts to show up in ChatGPT searches. Some parts of the community platform, like Vercel Community live sessions, run directly on Vercel with Next.js. We use Vercel Microfrontends to mount a Next.js app on the same domain as the Discourse app, for three reasons: To create new pages that would be impractical to imple

Chat SDK adds PostgreSQL state adapter

March 10, 2026

Chat SDK now supports PostgreSQL as a state backend, joining Redis and ioredis as a production-ready option with the new PostgreSQL adapter. Teams that already run PostgreSQL can persist subscriptions, distributed locks, and key-value cache state without adding Redis to their infrastructure. Try the PostgreSQL state adapter today: The adapter uses pg (node-postgres) with raw SQL queries and automatically creates the required tables on first connect. It supports TTL-based caching, distributed locking across multiple instances, and namespaced state via a configurable key prefix. Read the documentation to get started, browse the directory, or build your own adapter. Special thanks to @bai, whose community contribution in PR #154 laid the groundwork for this adapter. Read more

AI Gateway supports OpenAI's Responses API

March 6, 2026

OpenAI's Responses API is now available through AI Gateway. The Responses API is a modern alternative to the Chat Completions API. Point your OpenAI SDK to AI Gateway's base URL and use the creator/model names to route requests. TypeScript and Python are both supported. All of the functionality in the Responses API was already accessible through AI Gateway via the AI SDK and Chat Completions API, but you can now use the Responses API directly. What you can do Text generation and streaming: Send prompts, get responses, stream tokens as they arrive Tool calling: Define functions the model can invoke, then feed results back Structured output: Constrain responses to a JSON schema Reasoning: Control how much effort the model spends thinking with configurable effort levels Getting started Install the OpenAI SDK and point it at AI Gateway. Basic example: text generation Send a prompt and get a response from any supported model. Structured output with reasoning Combine reasoning levels with a JSON schema to get structured responses. To learn more about the Responses API, read the documentation. Read more

Chat SDK adds table rendering and streaming markdown

March 6, 2026

Chat SDK now renders tables natively across all platform adapters and converts markdown to each platform's native format during streaming. The Table() component is a new card element in Chat SDK that gives you a clean, composable API for rendering tables across every platform adapter. Pass in headers and rows, and Chat SDK handles the rest. The adapter layer converts the table to the best format each platform supports. Slack renders Block Kit table blocks, Teams and Discord use GFM markdown tables, Google Chat uses monospace text widgets, and Telegram converts tables to code blocks. GitHub and Linear already supported tables through their markdown pipelines and continue to work as before. Plain markdown tables (without Table()) are also converted through the same pipeline. Streaming markdown has also improved across the board. Slack's native streaming path now renders bold, italic, lists, and other formatting in real time as the response arrives, rather than resolving when the message is complete. All other platforms use the fallback streaming path, so streamed text now passes through each adapter's markdown-to-native conversion pipeline at each intermediate edit. Previously, these

v0 API now supports custom MCP servers

March 6, 2026

The v0 API now supports connecting to any custom MCP server. Teams can configure new servers programmatically by providing the necessary endpoint and authentication details. Once configured, you can make these custom servers available directly during a v0 chat session by referencing the server ID: Visit the v0 API docs. Read more

Skip unaffected builds for projects in Bun monorepos

March 6, 2026

Skipping unaffected builds in monorepos now detects Bun lockfiles, extending the same compatibility already available for other package managers. When Vercel evaluates which projects to build, it reads lockfile changes to determine whether dependencies have changed. Teams using Bun can now rely on this detection to skip builds for projects that haven't changed, reducing unnecessary build time across monorepos. See the monorepo documentation to learn how skipping unaffected projects works. Read more

Deployment step now 15% faster

March 6, 2026

Builds on Vercel now deploy 1.2 seconds faster on average, with more complex projects seeing the biggest gains (up to 3.7 seconds). The improvement comes from optimizing how credentials are provisioned during the build process, eliminating a blocking step that previously added latency at the end of every build. Learn more in the builds documentation. Read more

Create private blob stores with a single click in v0

March 5, 2026

Teams can now create private and public blob stores with a single click in v0. When adding Vercel Blob to a chat, a dialog lets you select your preferred region and access type. Private storage is selected by default and requires authentication to access sensitive files, while public storage allows direct reads for assets like media. Once connected, the agent automatically understands your store's configuration. It writes the correct implementation for your choice, setting up authenticated delivery routes for private stores or direct URLs for public ones, without requiring you to write any code manually. Learn more in the Vercel Blob documentation. Read more

GPT 5.4 is now on AI Gateway

March 5, 2026

GPT-5.4 and GPT-5.4 Pro are now available on AI Gateway. This model brings the agentic and reasoning leaps from GPT-5.3-Codex to all domains. This includes knowledge work like reports, spreadsheets, presentations, and analysis in addition to coding. It handles complex multi-step workflows more reliably, including tasks that involve tools, research, and pulling from multiple sources. GPT-5.4 is faster and also more token-efficient than previous iterations (GPT-5.2). GPT-5.4 Pro is for developers who need maximum performance on the most complex tasks. To use this model, set model to openai/gpt-5.4 or openai/gpt-5.4-pro in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

From idea to secure checkout in minutes with Stripe

March 5, 2026

Building commerce applications looks very different than it did even a few years months ago. Teams are no longer treating storefronts and billing systems as long-running integration projects that happen after the product is complete. They iterate quickly, deploy globally by default, and increasingly rely on AI tools to generate UI, checkout flows, and subscription logic. Commerce is becoming more programmable and increasingly agent-driven. As AI systems begin to generate storefronts, assemble checkout flows, and optimize billing logic, the setup, integrations, and infrastructure need to be just as composable and automated. With tools like v0, and agentic coding agents working with Vercel CLI and Vercel Marketplace, developers can move from idea to deployed product much faster than before. As that workflow becomes more automated and AI-native, the surrounding systems need to keep pace, and the developer's experience (which includes the agent's) needs as much focus as the end-user's. An improved developer experience Historically, moving from a Stripe Sandbox to accepting live payments required retrieving API keys, copying them into environment variables, and verifying configuration a

Customize timeouts for faster automatic failover on Vercel AI Gateway

March 5, 2026

AI Gateway now supports per-inference provider timeouts for faster failover than the provider default. If a provider doesn't start responding within your configured timeout, AI Gateway aborts the request and falls back to the next available provider. Provider timeouts are available in beta for BYOK (Bring Your Own Key) credentials only, with support for system provider timeouts coming soon. Note that some providers don't support stream cancellation, so you may still be charged for timed-out requests depending on the provider. Basic usage Set timeouts per provider in milliseconds using providerTimeouts in providerOptions.gateway. Advanced usage with multiple providers and failover Use with order to control both the provider sequence and failover speed. For more information, read the custom provider timeouts documentation. Read more

Vercel CDN gets a new dashboard experience

March 5, 2026

Vercel's CDN now has a new dashboard to give you a single place to track global traffic distribution and top-line CDN metrics, manage caching, and update routing rules. The new experience includes: Overview: A live map of your project's global traffic distribution across Vercel Regions and top-level request volume and cache performance metrics. Caches: A redesigned page for purging content across Vercel CDN's caching layers, which was previously under project settings. Project-level Routing: A new project-level UI for updating routing rules, like setting response headers or rewriting to an external API, without triggering a new deployment. Learn more about Vercel's CDN or visit the CDN tab for your project to see the updates. Read more

Vercel's CDN now supports updating routing rules without a new deployment

March 5, 2026

You can now create and update routing rules within a project, such as setting response headers or rewrites to an external API, without building a new deployment. Project-level routing rules are available via the dashboard, API, CLI, and Vercel SDK and take effect instantly after you make and publish the change. Project-level routes run after bulk redirects and before your deployment config's routes. With this addition, Vercel's CDN now supports three routing mechanisms: Routes defined in your deployment configuration (via vercel.json, vercel.ts, or next.config.js) Bulk redirects Project-level routes Project-level routes are available on all plans starting today. Read the documentation or go to the CDN tab in your project dashboard to get started. Read more

Streamdown 2.4: More customization, accessibility and custom rendering

March 5, 2026

Streamdown v2.4 introduces customization hooks, accessibility features, and user experience improvements for developers rendering markdown. Teams can now customize the appearance of their markdown output using several new properties. You can override the built-in icons by passing a specific component map to the icons prop. The createCodePlugin now accepts a themes option for light and dark Shiki themes, a startLine meta option for custom starting line numbers, and an inlineCode virtual component for styling inline code independently from blocks. Streamdown now supports internationalization and text direction. The dir prop automatically applies left-to-right or right-to-left formatting based on the first strong Unicode character, and the translations prop supports custom languages. Tables include a fullscreen overlay controlled via the controls prop, complete with scroll locking and Escape key support. Developers can hook into streaming events using the onAnimationStart and onAnimationEnd callbacks. This release fixes empty lines collapsing in syntax-highlighted blocks and prevents ordered lists from retriggering animations during streaming. For projects using Tailwind v4, the new p

Run cron jobs from deployment summary

March 5, 2026

You can now run cron jobs from your application in the summary section of your deployments dashboard. Try it out by deploying a Vercel cron job template. Once you deploy, Vercel automatically registers your cron jobs. Learn more in the cron jobs documentation. Read more

Stripe is now generally available on the Marketplace and v0

March 5, 2026

You can now connect your production Stripe account to Vercel and start accepting real payments. The integration securely provisions your API keys as environment variables and supports both sandbox and live modes. Test your payment flows in sandbox, then move to production without manually exchanging or managing keys. Built in collaboration with Stripe, the new key management APIs make it possible to reduce setup friction while strengthening security from day one. This unlocks real production use cases like: Live ecommerce: Accept real payments and manage checkout flows for production storefronts Production SaaS billing: Charge customers for subscriptions, usage, and invoices from day one Shipping to real users: Move from sandbox to production without re-wiring your integration Get started today with this example to build your first online simple store using Vercel and Stripe. See the documentation to learn more. Read more

MCP Apps support on Vercel

March 4, 2026

Teams can now build and deploy MCP Apps on Vercel with full support for Next.js. MCP Apps are similar to ChatGPT apps, but are a provider-agnostic open standard for embedded UIs. They run inside iframes and communicate with any compatible host, such as Cursor, Claude.ai, and ChatGPT, using a shared bridge. This architecture uses ui/* JSON-RPC over postMessage, enabling a single UI to function across any compatible host without platform-specific integrations. By combining this standard with Next.js on Vercel, developers can leverage Server-Side Rendering (SSR) and React Server Components to build portable, high-performance agent interfaces. Deploy the template or learn more in the documentation. Read more

Inception Mercury 2 is live on AI Gateway

March 4, 2026

Mercury 2 from Inception is now available on Vercel AI Gateway. Mercury 2 from Inception is now available on AI Gateway. Mercury 2 delivers reasoning-grade quality at real-time latency, making it ideal for agentic loops, coding assistants, voice interfaces, and RAG pipelines where compounding latency is the bottleneck. To use Mercury 2, set model to inception/mercury-2 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Vercel Workflow is now twice as fast

March 3, 2026

Server-side performance for Vercel Workflow, the fully managed platform built on top of the open-source Workflow Development Kit (WDK), is now twice as fast, delivering a 54% median improvement across the board. Over the last two weeks, the median API response time has been reduced from 37ms to 17ms, with queue latency, Time to First Byte (TTFB), and per-step overhead all reduced. Workflows that coordinate multiple steps benefit the most, as lower overhead compounds across each step in a run. To get these and future speedups, update to the latest version of the Workflow DevKit (workflow@4.1.0-beta.60 or newer) or view the documentation. Read more

GPT 5.3 Chat is now on AI Gateway

March 3, 2026

GPT-5.3 Chat (GPT 5.3 Instant) is now available on AI Gateway. This update focuses on tone, relevance, and conversational flow for more accurate answers, better-contextualized web results, and fewer unnecessary refusals and caveats. It also reduces hallucination rates and produces smoother and more direct responses. To use this model, set model to openai/gpt-5.3-chat in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Gemini 3.1 Flash Lite is now on AI Gateway

March 3, 2026

Gemini 3.1 Flash Lite from Google is now available on AI Gateway. This model outperforms 2.5 Flash Lite on overall quality, with notable improvements in translation, data extraction, and code completion. Gemini 3.1 Flash Lite is best suited for high-volume agentic tasks, data extraction, and applications where budget and latency are the primary evaluation constraints. To use this model, set model to google/gemini-3.1-flash-lite-preview in the AI SDK. This model supports four thinking levels, minimal, low, medium, and high. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Building Slack agents can be easy

March 3, 2026

Slack is already where teams work. It provides a natural interface for agents, with messages, threads, buttons, and events, so you don't need to invent a new UI or onboarding flow. Getting from "I want a Slack agent" to a running deployment, though, means coordinating across a lot of systems: Creating an app in the Slack API console Configuring OAuth scopes and event subscriptions Writing webhook handlers and signature verification Deploying to infrastructure that can handle Slack's 3-second response window Each piece has its own docs, and they all need to work together. Coding agents like Claude Code, OpenCode, Cursor, and GitHub Copilot are well suited for exactly this kind of coordination because they can read docs, reason through dependencies, and generate code in seconds. We built the Slack agent skill to take advantage of that. It builds on our Slack Agent Template, works with the coding agent of your choice, and takes you from idea to a deployed Slack agent on Vercel in a single session, automating steps when possible and showing you exactly what to click when it can't. From idea to deployed agent with the skill wizard Install the skill and run the wizard: Then run the skill

Scaling redirects to infinity on Vercel

March 3, 2026

Redirects are trivial at a small scale, but at millions, latency and cost become real systems problems. Previously on Vercel, redirects were handled by routing rules and middleware. Routing rules support up to 2,000 complex redirects with wildcards, and they function as an ordered list evaluated in sequence. Each rule may involve regex matching, meaning a single request could trigger many expensive evaluations. This is acceptable for a few thousand routing rules, but as counts grow, per-request work increases linearly. Middleware offers more flexibility, but it adds latency by running extra code on every request. To serve millions of redirects with low latency, we needed a dedicated lookup path with near-constant or logarithmic time per request. Building on our previous work to make global routing faster with Bloom filters, we found a way to scale to millions of redirects. What we optimized for Scale: Support millions of static redirects per project Runtime behavior: No additional latency cost for projects that don't configure redirects A fast "no redirect" path, since most requests won't be redirected Low process memory usage, relying on external storage and caching layers instead

Vercel Sandbox now accepts environment variables at creation

March 3, 2026

The Vercel Sandbox SDK and CLI now support setting environment variables at sandbox creation that are automatically available to every command. When running multi-step processes in a Vercel Sandbox like installing dependencies, building a project or starting a dev server, each step often needs the same environment variables. Now, these are available with every runCommand call. Environment variables passed to Sandbox.create() are inherited by all commands automatically. Per-command env in runCommand can still override individual values when needed. Update to the latest Sandbox CLI and SDK, run npm i @vercel/sandbox to get started. Read more

Advancing Python typing

March 2, 2026

We’re excited to share a year-long research effort aimed at making Python’s type system more expressive and composable, something closer in spirit to the programmable types in TypeScript, but carefully crafted for Python’s runtime model. The result is PEP 827: Type Manipulation. Python’s runtime is incredibly powerful: classes, methods, and even whole APIs can be generated on the fly from a few lines of code. Concepts like metaprogramming can transform class declarations, decorators can give functions and methods additional behaviors, and those are just a few examples. But Python's static typing often can’t “follow along” without typechecker plugins or boilerplate code. PEP 827 proposes a set of standard, type-level building blocks for introspecting existing types and constructing new ones, designed to help both type checkers and runtime tooling. FastAPI creator Sebastián Ramírez summed up the potential impact well on our post in the Python Discourse: Quick taste One concrete example is the familiar TypeScript utility types, like Pick and Omit. Here's Pick implemented in TypeScript and Python side by side: We can immediately see that the TypeScript dedicated typing syntax is short

Vercel CLI for Marketplace integrations optimized for agents

March 2, 2026

AI agents can now autonomously discover, install, and retrieve setup instructions for Vercel Marketplace integrations using the Vercel CLI. This lets agents configure databases, auth, logging, and other services end-to-end in one workflow. These capabilities are powered by the new discover and guide commands in the Vercel CLI. By using the --format=json flag with the discover command, the CLI provides non-interactive JSON output that benefits developers as well, making it easier to automate infrastructure, write custom scripts, and manage CI/CD pipelines. When building an application, agents begin by exploring available integrations using the discover command. After exploring the options, the agent can add an integration and then fetch getting started guides and code snippets for a specific integration using the guide command. The Vercel CLI returns this setup documentation in an agent-friendly markdown format. This allows the agent to easily parse the instructions, write the necessary integration code, and configure the project autonomously. For integrations with required metadata fields, agents can use the help command to determine the required inputs and pass them as options to

Gamma builds design-first agents with Vercel

February 28, 2026

Gamma began with a simple idea: what if your presentation could design itself? With a single sentence, users can generate a complete presentation that respects layout, spacing, and hierarchy. Columns reflow automatically. Diagrams adjust when new layers are added. The product handles the formatting so teams can stay focused on the ideas. That philosophy reflects the company's DNA. Of Gamma's first ten hires, three were designers. "The attention to detail and value placed on design has been baked into the culture from the very, very beginning," says Sherwin Yu, Head of AI and Product Engineering. "Our designers at Gamma are fantastic. They ship code, they're technical. They'll push to production." "There's a lot of discussion about how do we, whenever possible, elevate the user experience," Sherwin says. As adoption grew, the team realized generation was only the beginning. Real presentation work happens in iteration. Teams outline, restructure, refine tone, and polish visuals. In October 2025, Gamma launched Gamma Agent, a conversational editing that shifted the AI capabilities dramatically. Evolving complex agent architectures with AI SDK The first version of Gamma generated decks

How Avalara turns pipe dreams into patent-pending with v0

February 28, 2026

Avalara connects businesses to more than 1,400 systems to automate tax compliance around the world. It’s a massively complex ecosystem that spans ERP systems, finance platforms, and compliance tools, all talking to each other. For Chief Strategy and Product Officer Jayme Fishman, the path forward is modernizing how Avalara builds. His mandate is to drive digital transformation, with a sharp focus on AI and innovation. Enter Vercel’s v0, which translates plain language into working prototypes. Within months, the team built two new patent pending products—and along the way, changed how the company builds. Seeing is believing Before v0, bringing an idea to life required a mountain of slides, careful specs, and ample interpretation. Fishman might have a strong vision, but getting started meant writing everything down, then waiting for designers and engineers to bring it to life. “It could be a significant delay before we even had a conceptual mock-up.” That changed overnight. One of Avalara's biggest challenges was supporting customers who could be plugging into more than a thousand different systems. "We could provide technical documentation and show customers what to do," Fishman sai

Keeping community human while scaling with agents

February 27, 2026

At Vercel, our developer community is at the heart of everything we do. It's how we stay closest to the people using what we build. As our community grew, automation helped us scale. But questions still got lost, routing took time, and context switching pulled us away from the work that actually required our expertise. And automation could never help with the things that mattered most, the moments where you really connect with someone and help them. You can't use AI to replicate the feeling of talking to a person who really cares. So we built agents to take over the routing, triage, and follow-ups that don't need a human. We call this the Community Guardian. Let's talk about what it does, how we built it, and how anyone, including non-engineers, can ship agents too. The Community Guardian operations layer When a new post comes in, the Guardian analyzes it, checks for duplicates, and assigns it to the team member with the right specialty and bandwidth. Each person handles up to 10 questions before new ones go to someone else, keeping the workload balanced across time zones. Nothing gets overlooked. If a question goes unanswered for 48 hours, the Guardian reassigns it. It sends remin

Vercel Queues now in public beta

February 27, 2026

Vercel Queues is a durable event streaming system built with Fluid compute, and is now available in public beta for all teams. Vercel Queues also powers Workflow: use Queues for direct message publishing and consumption, Workflow for ergonomic multi step orchestration. Functions need a reliable way to defer expensive work and guarantee that tasks complete even when functions crash or new deployments roll out. Queues makes it simple to process messages asynchronously with automatic retries and delivery guarantees, providing at-least-once delivery semantics. How it works: Messages are sent to a durable topic The queue fans messages out to subscribed consumer groups. Each consumer group processes messages independently. The queue redelivers messages to consumer groups until successfully processed or expired. Publish messages from any route handler: Create a consumer: Configure the consumer group: Adding a trigger makes the route private: it has no public URL and only Vercel's queue infrastructure can invoke it. Vercel Queues is billed per API operation, starting at $0.60 per 1M operations, and includes: Multiple AZ synchronous replication At-least-once delivery Customizable visibility

Chat SDK adds Telegram adapter support

February 27, 2026

Chat SDK now supports Telegram, extending its single-codebase approach to Slack, Discord, GitHub, and Teams, with the new Telegram adapter. Teams can build bots that support mentions, message reactions, direct messages, and typing indicators. The adapter handles single file uploads and renders basic text cards, with buttons and link buttons that display as inline keyboard elements, allowing developers to create interactive workflows directly within Telegram chats. Get start with Telegram adapter setup: Telegram does not expose full historical message APIs to bots, so message history relies on adapter-level caching. Additionally, callback data is limited to 64 bytes, and the platform does not currently support modals or ephemeral messages. Read the documentation to get started, browse the directory, or build your own adapter. Read more

Developer role now available for Pro teams

February 26, 2026

Pro teams can now assign the Developer role to their members. Previously only available for Enterprise teams, the Developer role gives Pro teams more granular access control. Developers can safely deploy to projects on a team, with more limited team-wide configuration control and environment variables visibility. Owners can assign the Developer role to any existing seat or invite new members from the team members settings. Learn more about team level roles. Read more

New dashboard redesign is now the default

February 26, 2026

The new dashboard navigation is now the default experience for all Vercel users. Following a successful opt-in beta release in January it has now rolled out fully as of February 26, 2026, with several improvements made based on feedback. The redesigned navigation includes: New sidebar with horizontal tabs moved to a resizable sidebar that can be hidden when not needed Consistent tabs for unified navigations across both team and project levels Improved order with navigation items prioritized the most common developer workflows Projects as filters so you can switch between team and project versions of the same page in one click Optimized for mobile with floating bottom bar optimized for one-handed use No action is required. The new navigation is available to all users automatically. Open your dashboard to see the updated experience. Read more

Nano Banana 2 is live on AI Gateway

February 26, 2026

Gemini 3.1 Flash Image Preview (Nano Banana 2) is now available on AI Gateway. This release improves visual quality while maintaining the generation speed and cost of flash-tier models. Nano Banana 2 can use Google Image Search to ground outputs in real-world imagery. This helps with rendering lesser-known landmarks and objects by retrieving live visual data. This model also introduces configurable thinking levels (Minimal and High) to let the model reason through complex prompts before rendering. New resolutions and new aspect ratios (512p, 1:4 and 1:8) are available alongside the existing options to expand to support more types of creative assets. To use this model, set model to google/gemini-3.1-flash-image-preview in the AI SDK. Nano Banana 2 is a multimodal model. Use `streamText` or `generateText` to generate images alongside text responses. This example shows how the model can use web search to find live data. You can also change the thinking level: in this example, thinking is set to high for a more thorough response. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-tha

Activity log now tracks 100% of team and project changes

February 25, 2026

The activity log now captures every change made to your team and project settings, giving you complete visibility into who changed what and when. Previously, some settings changes went untracked. With 88 new events added, activity log coverage is now 100%, so no action goes unrecorded. Try it out or learn more about the activity log. Read more

How OpenEvidence built a healthcare AI that physicians actually trust

February 25, 2026

Andy Yoon was scrolling through Slack when he saw the message: OpenEvidence had gone viral on TikTok. Not "gaining traction.” Actually viral, reaching around two million views in less than a week. This is usually when you rally the troops, spin up emergency capacity, and start making phone calls you really didn't want to make. Andy, Lead Frontend Engineer, did none of those things. Instead, he watched the numbers climb. He checked the logs—everything green. Response times: still fast. Error rates: still near zero. Then he went back to whatever he was doing before, because there was nothing to fix. "Vercel has just completely scaled with that usage," he says. "We've never had it fall over due to capacity or had to provision anything extra. Just being able to trust that it's there, to the point where we don't really even think about it, is amazing." It was proof that they'd solved a problem most healthcare tech companies haven't figured out yet: how to move at startup speed while meeting hospital-grade reliability standards. When failure isn't an option The stakes are different for companies like OpenEvidence. If their product fails, it could result in someone making a bad medical de

Security boundaries in agentic architectures

February 24, 2026

Most agents today run generated code with full access to your secrets. As more agents adopt coding agent patterns, where they read filesystems, run shell commands, and generate code, they're becoming multi-component systems that each need a different level of trust. While most teams run all of these components in a single security context, because that's how the default tooling works, we recommend thinking about these security boundaries differently. Below we walk through: The actors in agentic systems Where security boundaries should go between them An architecture for running agent and generated code in separate contexts All agents are starting to look like coding agents More agents are adopting the coding agent architecture. These agents read and write to a filesystem. They run bash, Python, or similar programs to explore their environment. And increasingly, agents generate code to solve particular problems. Even agents that aren't marketed as "coding agents" use code generation as their most flexible tool. A customer support agent that generates and runs SQL to look up account data is using the same pattern, just pointed at a database instead of a filesystem. An agent that can

Python Vercel Functions bundle size limit increased to 500MB

February 24, 2026

The bundle size limit for Vercel Functions using the Python runtime is now 500MB, increasing the maximum uncompressed deployment bundle size from 250MB. (This 500MB limit is the standard path; Large Functions support Python bundles up to 5GB on Fluid compute.) Learn more in the functions limitations documentation, or deploy FastAPI or Flask on Vercel to get started. Read more

GPT 5.3 Codex is now on AI Gateway

February 24, 2026

GPT 5.3 Codex is now available on AI Gateway. GPT 5.3 Codex brings together the coding strengths of GPT-5.2-Codex and the reasoning depth of GPT-5.2 in a single model that's 25% faster and more token-efficient. Built for long-running agentic work, the model handles research, tool use, and multi-step execution across the full software lifecycle, from debugging and deployment to product documents and data analysis. Additionally, you can steer it mid-task without losing context. For web development, it better understands underspecified prompts and defaults to more functional, production-ready output. To use this model, set model to openai/gpt-5.3-codex in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Slack Agent Skill simplifies building Slack agents with coding assistants

February 24, 2026

The Slack Agent Skill is now available, enabling developers to build and deploy Slack agents in a single session with their coding agent of choice. The skill handles the complexity of OAuth configuration, webhook handlers, event subscriptions, and deployment so you can focus on what your agent should do rather than on infrastructure setup. The wizard walks through five stages: Project setup: Choose your LLM provider and initialize from the Slack Agent Template Slack app creation: Generate a customized app manifest and create the app in Slack's console Environment configuration: Set up signing secrets, bot tokens, and API keys with validation Local testing: Run locally with ngrok and verify the integration Production deployment: Deploy to Vercel with environment variables configured automatically Install the skill and run the wizard by invoking it in your coding agent (for example, /slack-agent new in Claude Code). Try the skill to make your custom agent or use the Slack Agent Template to deploy right away and customize later. Read more

Support for now.json will be removed on March 31, 2026

February 23, 2026

Support for the legacy now.json config file will be officially removed on March 31st, 2026. Migrate existing now.json files by renaming them to vercel.json, no other content changes are required. For more advanced use cases, try vercel.ts for programmatic project configuration. Learn more about configuring projects with vercel.json in the documentation. Read more

Introducing npm i chat – One codebase, every chat platform

February 23, 2026

Building chatbots across multiple platforms traditionally requires maintaining separate codebases and handling individual platform APIs. Today, we're open sourcing the new Chat SDK in public beta. It's a unified TypeScript library that lets teams write bot logic once and deploy it to Slack, Microsoft Teams, Google Chat, Discord, GitHub, and Linear. The event-driven architecture includes type-safe handlers for mentions, messages, reactions, button clicks, and slash commands. Teams can build user interfaces using JSX cards and modals that render natively on each platform. The SDK handles distributed state management using pluggable adapters for Redis, ioredis, and in-memory storage. You can post messages to any provider with strings, objects, ASTs and even JSX! Chat SDK post() functions accept an AI SDK text stream, enabling real-time streaming of AI responses and other incremental content to chat platforms. The framework starts with the core chat package and scales through modular platform adapters. Guides are available for building a Slack bot with Next.js and Redis, a Discord support bot with Nuxt, a GitHub bot with Hono, and automated code review bots. Explore the documentation t

Safely inject credentials in HTTP headers with Vercel Sandbox

February 23, 2026

Vercel Sandbox can now automatically inject HTTP headers into outbound requests from sandboxed code. This keeps API keys and tokens safely outside the sandbox VM boundary, so apps running inside the sandbox can call authenticated services without ever accessing the credentials. Header injection is configured as part of the network policy using transform. When the sandbox makes an HTTPS request to a matching domain, the firewall adds or replaces the specified headers before forwarding the request. This is designed for AI agent workflows where prompt injection is a real threat. Even if an agent is compromised, there's nothing to exfiltrate, as the credentials only exist in a layer outside the VM. Injection rules work with all egress network policy configurations, including open internet access. To allow general traffic while injecting credentials for specific services: Live updates Like all network policy settings, injection rules can be updated on a running sandbox without restarting it. This enables multi-phase workflows, inject credentials during setup, then remove them before running untrusted code: Key highlights Header overwrite: Injection applies to HTTP headers on outbound re

Skills Night: 69,000+ ways agents are getting smarter

February 20, 2026

The room was full of people who had already used skills. Tuesday night we hosted Skills Night in San Francisco, an event for developers building on and around skills.sh, the open skills ecosystem we've been growing since the idea started as a single weekend of writing. What began as Shu Ding sitting down to document everything he knows about React has grown into over 69,000 skills, 2 million skill CLI installs, and a community moving incredibly fast. Here is what we learned. Where this came from The origin story is worth retelling because it shapes how we think about the project. Shu Ding is one of the most talented web engineers I've ever worked with. He knows things about React and the browser that most people will never discover. Last year, he sat down on a weekend and wrote it all down. A kind of web bible. We wanted to figure out how to ship it. We considered a blog post or documentation that the next generation of models might eventually learn - but we wouldn't see the results until Claude Sonnet 8, or GPT-9. On the other hand, an MCP server felt too heavy for what was essentially a collection of markdown documents. Skills made sense as the quickest way to deliver on-demand k

Access billing usage and cost data via API

February 19, 2026

Vercel now supports programmatic access to billing usage and cost data through the API and CLI. The new /billing/charges endpoint returns data in the FOCUS v1.3 open-standard format, allowing teams to ingest cost data into FinOps tools without custom transformation logic. The endpoint supports 1-day granularity with a maximum date range of one year. Responses are streamed as newline-delimited JSON (JSONL) to handle large datasets efficiently. SDK usage curl usage CLI usage For quick introspection, the vercel usage command displays billing usage for the current period or a custom date range. This includes credit-use and costs for each service. View usage for the current billing period View usage for a custom date range Vantage has also released a native integration that connects Vercel teams to Vantage accounts. This automatically syncs usage and cost data alongside other tools, simplifying cost observability. Read the Vantage announcement blog for details. Learn more in the API documentation and CLI reference. Read more

Streamdown 2.3 — Refreshed design and interactive playground

February 19, 2026

Streamdown 2.3 enhances design consistency by applying a unified nested-card design to tables, code blocks, and Mermaid diagrams. Action buttons now remain sticky during scroll, and code blocks render plain text immediately to reduce perceived latency before syntax highlighting loads. To accelerate testing, the new interactive playground supports real-time execution with custom markdown and editable props. This enables faster experimentation with configuration changes without spinning up a local project. New hooks and utilities provide improved control over rendering. The useIsCodeFenceIncomplete hook detects in-progress fenced code blocks during streaming. Tables now support copying as Markdown, and a new HTML indentation normalization property handles inconsistent whitespace in raw input. Image rendering also includes improved error handling with custom messaging. Documentation has been reorganized for easier reference. Plugin documentation for CJK, Math, and Mermaid is now consolidated into dedicated pages, and the redesigned homepage links directly to templates for faster onboarding. This release also resolves issues with nested HTML block parsing, custom tag handling, Mermaid

Gemini 3.1 Pro is live on AI Gateway

February 19, 2026

Gemini 3.1 Pro Preview from Google is now available on AI Gateway. This model release brings quality improvements across software engineering and agentic workflows, with enhanced usability for real-world tasks in finance and spreadsheet applications. Gemini 3.1 Pro Preview introduces more efficient thinking across use cases, reducing token consumption while maintaining performance. To use this model, set model to google/gemini-3.1-pro-preview in the AI SDK. This model supports the medium thinking level for finer control over the trade-offs between cost, performance, and speed. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Private storage for Vercel Blob, now available in public beta

February 19, 2026

Update (June 30, 2026): Vercel Private Blob is now generally available. See the GA changelog for details. Vercel Blob now supports private storage for sensitive files like contracts, invoices, and internal reports. Private storage requires authentication for all operations, preventing exposure via public URLs. Public storage allows public reads for media assets, while private storage requires authentication. Create a private store via the Storage dashboard or with the CLI: CLI command When created inside a linked Vercel project, the CLI prompts you to connect the store, automatically adding the BLOB_READ_WRITE_TOKEN environment variable. The SDK uses this variable to authenticate operations in your deployments. SDK installation To upload, use put or upload with the access: 'private' option. Upload example To download, use the get method to stream files. Retrieval example Private storage is in beta on all plans with standard Vercel Blob pricing. Learn more about private storage. Read more

Video Generation with AI Gateway

February 19, 2026

AI Gateway now supports video generation, so you can create cinematic videos with photorealistic quality, synchronized audio, generate personalized content with consistent identity, all through AI SDK 6. Two ways to get started Video generation is in beta and currently available for Pro and Enterprise plans and paid AI Gateway users. AI SDK 6: Generate videos programmatically with the same interface you use for text and images. One API, one authentication flow, one observability dashboard across your entire AI pipeline. AI Gateway Playground: Experiment with video models with no code in the configurable AI Gateway playground that's embedded in each model page. Compare providers, tweak prompts, and download results without writing code. To access, click any video gen model in the model list. Four initial video models; 17 variations Grok Imagine from xAI is fast and great at instruction following. Create and edit videos with style transfer, all in seconds. Wan from Alibaba specializes in reference-based generation and multi-shot storytelling, with the ability to preserve identity across scenes. Kling excels at image to video and native audio. The new 3.0 models support multishot vide

Grok Imagine Video on AI Gateway

February 19, 2026

Generate high-quality videos with natural motion and audio using xAI's Grok Imagine Video, now in AI Gateway. Try it out now via the v0 Grok Creative Studio, AI SDK 6 or by selecting the model in the AI Gateway playground. Grok Imagine is known for realistic motion and strong instruction following: Fast Generation: Generates clips in seconds rather than minutes Instruction Following: Understands complex prompts and follow-up instructions to tweak scenes Video Editing: Transform existing videos by changing style, swapping objects, or altering scenes Audio & Dialogue: Native audio generation with natural, expressive voices and accurate lip-sync Three ways to get started Video generation is in beta and currently available for Pro and Enterprise plans and paid AI Gateway users. v0 Grok Creative Studio: The v0 team created a template that is powered by AI Gateway to create and showcase Grok Video and Image generations. AI SDK 6: Generate videos programmatically AI SDK 6's generateVideo. Gateway Playground: Experiment with video models with no code in the configurable AI Gateway playground that's embedded in each model page. Compare providers, tweak prompts, and download results without

Wan models on AI Gateway

February 19, 2026

Generate stylized videos and transform existing footage with Alibaba's Wan models, now available through AI Gateway. Try them out now via AI SDK 6 or by selecting the models in the AI Gateway playground. Wan produces artistic videos with smooth motion and can use existing content to keep videos consistent: Character Reference (R2V): Extract character appearance and voice from reference videos/images to generate new scenes Flash Variants: Faster generation times for quick iterations Flexible Resolutions: Support for 480p, 720p, and 1080p output Two ways to get started Video generation is in beta and currently available for Pro and Enterprise plans and paid AI Gateway users. AI SDK 6: Generate videos programmatically AI SDK 6's generateVideo. Gateway Playground: Experiment with video models with no code in the configurable AI Gateway playground that's embedded in each model page. Compare providers, tweak prompts, and download results without writing code. To access, click any video gen model in the model list. Available Models Model Type Description alibaba/wan-v2.6-t2v Text-to-Video Generate videos from text prompts alibaba/wan-v2.6-i2v Image-to-Video Animate still images alibaba/wa

Kling video models on AI Gateway

February 19, 2026

Kling video models are now available in AI Gateway, including the newest Kling 3.0 models. Generate cinematic videos from text, images, or motion references with Kling's state-of-the-art video models, now available through AI Gateway and AI SDK. Kling models are known for their image to video models and multishot capabilities: Image-to-Video Capabilities: Strong at animating still images into video clips Realistic Motion and Physics: Known for coherent motion, facial expressions, and physical interactions High Resolution Output: Supports up to 1080p generation (pro mode) Multishot Narratives: Kling 3.0 can generate multi-scene videos from a single narrative prompt Audio Generation: Create synchronized sound effects and ambient audio alongside your video First & Last Frame Control: Specify both start and end frames for precise scene transitions Two ways to get started Video generation is in beta and currently available for Pro and Enterprise plans and paid AI Gateway users. AI SDK 6: Generate videos programmatically AI SDK 6's generateVideo. Gateway Playground: Experiment with video models with no code in the configurable AI Gateway playground that's embedded in each model page. Com

Veo video models on AI Gateway

February 19, 2026

Generate photorealistic videos with synchronized audio using Google's Veo models, now available through AI Gateway. Try them out now via AI SDK 6 or by selecting the models in the AI Gateway playground. Veo models are known for their cinematic quality and audio generation: Native Audio Generation: Automatically generate realistic sound effects, ambient audio, and even dialogue that matches your video Up to 1080p Resolution: Generate videos at 720p and 1080p Photorealistic Quality: Realism for nature, wildlife, and cinematic scenes Image-to-Video: Animate still photos with natural motion Fast Mode: Quicker generation when you need rapid iterations Two ways to get started Video generation is in beta and currently available for Pro and Enterprise plans and paid AI Gateway users. AI SDK 6: Generate videos programmatically AI SDK 6's generateVideo. Gateway Playground: Experiment with video models with no code in the configurable AI Gateway playground that's embedded in each model page. Compare providers, tweak prompts, and download results without writing code. To access, click any video gen model in the model list. Available Models Model Description google/veo-3.1-generate-001 Latest g

We Ralph Wiggumed WebStreams to make them 10x faster

February 18, 2026

When we started profiling Next.js server rendering earlier this year, one thing kept showing up in the flamegraphs: WebStreams. Not the application code running inside them, but the streams themselves. The Promise chains, the per-chunk object allocations, the microtask queue hops. After Theo Browne's server rendering benchmarks highlighted how much compute time goes into framework overhead, we started looking at where that time actually goes. A lot of it was in streams. Turns out that WebStreams have an incredibly complete test suite, and that makes them a great candidate for doing an AI-based re-implementation in a purely test-driven and benchmark-driven fashion. This post is about the performance work we did, what we learned, and how this work is already making its way into Node.js itself through Matteo Collina's upstream PR. The problem Node.js has two streaming APIs. The older one (stream.Readable, stream.Writable, stream.Transform) has been around for over a decade and is heavily optimized. Data moves through C++ internals. Backpressure is a boolean. Piping is a single function call. The newer one is the WHATWG Streams API: ReadableStream, WritableStream, TransformStream. This

Redesigned search and filtering for runtime logs

February 18, 2026

The Runtime Logs search bar in your project dashboard has been redesigned to make filtering and exploring your logs faster and more intuitive. Structured filters. When you type a filter like level:error or status:500, the search bar parses it into a visual pill you can read at a glance and remove with a click. Complex queries with multiple filters become easy to scan and edit without retyping anything Smarter suggestions. As you type, the search bar suggests filter values based on your actual log data. Recent queries are saved per-project and appear at the top, so you can rerun common searches without retyping them Better input handling. The search bar validates your filters as you type and flags errors with a tooltip so you can fix typos before running a search. Pasting a Vercel Request ID automatically converts it into a filter These improvements are available now in your project dashboard. Learn more about runtime logs. Read more

How Stably ships AI testing agents in hours, not weeks

February 17, 2026

How the 6-person team at Stably ships AI testing agents faster with Vercel—moving from weeks to hours. Their shift highlights how Vercel's platform eliminates infrastructure anxiety, boosting autonomous testing and enabling quick enterprise growth. Read more

Automatic build fix suggestions with Vercel Agent

February 17, 2026

You can now get automatic code-fix suggestions for broken builds from the Vercel Agent, directly in GitHub pull request reviews or in the Vercel Dashboard. When the Vercel Agent reviews your pull request, it now scans your deployments for build errors, and when it detects failures it automatically suggests a code fix based on your code and build logs. In addition, Vercel Agent can automatically suggest code fixes inside the Vercel dashboard whenever a build error is detected, and suggests a code change to a GitHub Pull Request for review before merging with your code. Get started with Vercel Agent code review in the Agent dashboard, or learn more in the documentation. Read more

Automated security audits now available for skills.sh

February 17, 2026

Skills on the skills.sh now have automated security audits to help developers use skills with confidence. Working with our partners Gen, Socket, and Snyk, these independent security reports allow us to rapidly scale and audit over 60,000 skills and counting. Skills.sh provides greater ecosystem support with: Transparent results: Security audits appear publicly on each skill's detail page. Leaderboard protection : Skills flagged as malicious are automatically hidden from the leaderboard and search results. If you navigate directly to a flagged skill, a warning note appears before installation. Security validation: As of skills@1.4.0, adding skills clearly displays audit results and risk levels before installation. Learn more at skills.sh. Read more

Recraft V4 on AI Gateway

February 17, 2026

Recraft V4 is now available on AI Gateway. A text-to-image model built for professional design and marketing use cases, V4 was developed with input from working designers. The model has improvements with photorealism, with realistic skin, natural textures, and fewer synthetic artifacts. It also produces images with clean lighting and varied composition. For illustration, the model can generate original characters with less predictable color palettes. There are 2 versions: V4: Faster and more cost-efficient, suited for everyday work and iteration V4 Pro: Generates higher-resolution images for print-ready assets and large-scale use To use this model, set model to recraft/recraft-v4-pro or recraft/recraft-v4 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Vercel Sandbox snapshots now allow custom retention periods

February 17, 2026

Snapshots created with Vercel Sandbox now have configurable expiration, instead of the previous 7 days limit, along with higher defaults. The expiration can be configured between 1 day to infinity. If not provided, the default snapshot expiration is 30 days. You can also configure this in the CLI. Read the documentation to learn more about snapshots. Read more

Claude Sonnet 4.6 is live on AI Gateway

February 17, 2026

Claude Sonnet 4.6 from Anthropic is now available on AI Gateway with the 1M token context window. Sonnet 4.6 approaches Opus-level intelligence with strong improvements in agentic coding, code review, frontend UI quality, and computer use accuracy. The model proactively executes tasks, delegates to subagents, and parallelizes tool calls, with MCP support for scaled tool use. As a hybrid reasoning model, Sonnet 4.6 delivers both near-instant responses and extended thinking within the same model. To use this model, set model to anthropic/claude-sonnet-4.6 in the AI SDK. This model supports effort and thinking type adaptive: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Improved streaming runtime logs exports

February 17, 2026

With runtime logs, you can view and export your logs. Exports now stream directly to the browser - your download starts immediately and you can continue to use the Vercel dashboard while the export runs in the background. This eliminates the need to wait for large files to buffer. Additionally, we've added two new options: You can now export exactly what's on your screen or all requests matching your current search. All plans can export up to 10,000 requests per export, and Observability Plus subscribers can export up to 100,000 requests. Exported log data is now indexed by request to ensure consistency with the Runtime Logs dashboard interface. Export limits are now applied by request to ensure that the exported data matches the filtered requests shown on the dashboard. Learn more about runtime logs. Read more

How Ramp kept 100% uptime through 100x traffic surges on Vercel

February 17, 2026

Ramp on Vercel Full site migration to Next.js on Vercel 100% uptime and 0% error rates through 100x surges AI SDK implemented to build semantic search across scattered internal assets Ramp is a financial operations platform that combines bill payments, corporate cards, procurement, and travel to help businesses save time and money. The team rebuilt its marketing site on Next.js and Vercel, migrating its highest-traffic and highest-conversion pages one at a time and A/B testing each against Core Web Vitals. The site has held 100% uptime through traffic surges of 100x during stunt-marketing events, and the team is now using the Vercel AI SDK to build semantic search across disparate internal assets. Read more

Qwen 3.5 Plus is on AI Gateway

February 16, 2026

Qwen 3.5 Plus is now available on AI Gateway. The model comes with a 1M context window and built-in adaptive tool use. Qwen 3.5 Plus excels at agentic workflows, thinking, searching, and using tools across multimodal contexts, making it well-suited for web development, frontend tasks, and turning instructions into working code. Compared to Qwen 3 VL, it delivers stronger performance in scientific problem solving and visual reasoning tasks. To use this model, set model to alibaba/qwen3.5-plus in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Stale-if-error cache-control directive now supported for all responses

February 13, 2026

Vercel CDN now supports the stale-if-error directive with Cache-Control headers, enabling more resilient caching behavior during origin failures. You can now use the stale-if-error directive to specify how long (in seconds) a stale cached response can still be served if a request to the origin fails. When this directive is present and the origin returns an error, the CDN may serve a previously cached response instead of returning the error to the client. Stale responses may be served for errors like 500 Internal Server Errors, network failures, or DNS errors. This allows applications to remain available and respond gracefully when upstream services are temporarily unavailable. Read the stale-if-error documentation to learn more. Read more

Browserbase joins the Vercel Agent Marketplace

February 12, 2026

Browserbase is now available on the Vercel Marketplace, allowing teams to run browser automation for AI agents without managing infrastructure. This integration connects agents to remote browsers over the Chrome DevTools Protocol (CDP), enabling workflows that require interacting with real websites, such as signing in to dashboards, filling out forms, or navigating dynamic pages. With this one-click integration, teams benefit from unified billing and infrastructure designed for long-lived, stateful sessions. Key capabilities include: Install and connect with a single API key Connect agents to remote browsers over CDP Reduce operational complexity for browser-based agent workflows Work with Vercel Sandbox and AI Gateway Also available today is support for Web Bot Auth for Browserbase, enabling agents to reliably browse Vercel-hosted deployments without interruption from security layers. Get started with Browserbase on the Vercel Marketplace or try this example to see it in action. Read more

Use MiniMax M2.5 on AI Gateway

February 12, 2026

MiniMax M2.5 is now available on AI Gateway. M2.5 plans before it builds, breaking down functions, structure, and UI design before writing code. It handles full-stack projects across Web, Android, iOS, Windows, and Mac, covering the entire development lifecycle from initial system design through code review. Compared to M2.1, it adapts better to unfamiliar codebases and uses fewer search rounds to solve problems. To use this model, set model to minimax/minimax-m2.5 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

New deployments with vulnerable versions of the third-party package next-mdx-remote are now blocked by default

February 12, 2026

Any new deployment containing a version of the third-party package next-mdx-remote that is vulnerable to CVE-2026-0969 will now automatically fail to deploy on Vercel. We strongly recommend upgrading to a patched version regardless of your hosting provider. This automatic protection can be disabled by setting the DANGEROUSLY_DEPLOY_VULNERABLE_CVE_2026_0969=1 environment variable on your Vercel project. Learn more Read more

Vercel Flags is now in public beta

February 11, 2026

Vercel Flags is a feature flag provider built into the Vercel platform. It lets you create and manage feature flags with targeting rules, user segments, and environment controls directly in the Vercel Dashboard. The Flags SDK provides a framework-native way to define and use these flags within Next.js and SvelteKit applications, integrating directly with your existing codebase: And you can use them within your pages like: For teams using other frameworks or custom backends, the Vercel Flags adapter supports the OpenFeature standard, allowing you to combine feature flags across various systems and maintain consistency in your flag management approach: Vercel Flags is priced at $30 per 1 million flag requests ($0.00003 per event), where a flag request is any request to your application that reads the underlying flags configuration. A single request evaluating multiple feature flags of the same source project still counts as one flag request. Vercel Flags is now in beta and available to teams on all plans. Learn more about Vercel Flags to get started with feature flag management. Read more

GLM-5 is live on AI Gateway

February 11, 2026

You can now access GLM-5 via AI Gateway with no other provider accounts required. GLM-5 from Z.AI is now available on AI Gateway. Compared to GLM-4.7, GLM-5 adds multiple thinking modes, improved long-range planning and memory, and better handling of complex multi-step agent tasks. It's particularly strong at agentic coding, autonomous tool use, and extracting structured data from documents like contracts and financial reports. To use this model, set model to zai/glm-5 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Advanced egress firewall filtering for Vercel Sandbox

February 11, 2026

Vercel Sandbox can now enforce egress network policies through Server Name Indication (SNI) filtering and CIDR blocks, giving you control over which hosts a sandbox can reach. Outbound TLS connections are matched against your policy at the handshake, unauthorized destinations are rejected before any data is transmitted. By default, sandboxes have unrestricted internet access. When running untrusted or AI generated code, you can lock down the network to only the services your workload actually needs. A compromised or hallucinated code snippet cannot exfiltrate data or make unintended API calls, traffic to any domain not on your allowlist is blocked. Going beyond IP based rules to host based The modern internet runs on hostnames, not IP addresses, a handful of addresses serve thousands of domains. Traditional IP-based firewall rules can't precisely distinguish between them. Host-based egress control typically requires an HTTP proxy, but that breaks non-HTTP protocols like Redis and Postgres. Instead, we built an SNI-peeking firewall that inspects the initial unencrypted bytes of a TLS handshake to extract the target hostname. Since nearly all internet traffic is TLS-encrypted today,

Support for Sign in with Apple

February 10, 2026

The login experience now supports Sign in with Apple, enabling faster access for users with Apple accounts. If your Apple account uses an Apple email (@icloud.com, @mac.com, @me.com, etc.) that matches your Vercel account's email, you can use the Apple button from the login screen and your accounts will be automatically linked. If the emails don't match, you can manually connect your Apple account from your account settings once logged in. Read more

vercel logs CLI command now optimized for agents with historical log querying

February 10, 2026

The vercel logs command has been rebuilt with more powerful querying capabilities, designed with agent workflows in mind. You can now query historical logs across your projects and filter by specific criteria, such as project, deploymentID, requestID, and arbitrary strings, to find exactly what you need. The updated command uses git context by default, automatically scoping logs to your current repository when run from a project directory. This makes it easy to debug issues during development without manually specifying project details. Whether you're debugging a production issue or building automated monitoring workflows, the enhanced filtering gives you precise control over log retrieval across your Vercel projects. Learn about Vercel CLI and vercel logs command. Read more

Agents can now access runtime logs with Vercel's MCP server

February 10, 2026

Agents can now access runtime logs through Vercel's MCP server. The get_runtime_logs tool lets agents retrieve Runtime Logs for a project or deployment. Runtime logs include logs generated by Vercel Functions invocations in preview and production deployments, including function output and console.log messages. This enables agents to: debug failing requests inspect function output search logs for specific errors or request IDs investigate runtime behavior across deployments Get started with the Vercel MCP server. Read more

PostHog joins the Vercel Marketplace

February 10, 2026

PostHog is now available in the Vercel Marketplace as a feature flags, experimentation and Analytics provider. With this integration, you can now: Declare flags in code using Flags SDK and the @flags-sdk/posthog adapter Toggle features in real time for specific users or cohorts Roll out changes gradually using percentage-based rollouts Run A/B tests to validate impact before a full release This integration helps teams building on Vercel ship with more confidence. You can test in production, reduce release risk, and make data-driven decisions based on real user behavior, all within your existing Vercel workflows. Create a flags.ts file with an identify function and a flag check: Check out the PostHog template to learn more about this integration. Read more

Anyone can build agents, but it takes a platform to run them

February 9, 2026

Prototyping is democratized, but production deployment isn't. AI models have commoditized code and agent generation, making it possible for anyone to build sophisticated software in minutes. Claude can scaffold a fully functional agent before your morning coffee gets cold. But that same AI will happily architect a $5,000/month DevOps setup when the system could run efficiently at $500/month. In a world where anyone can build internal tools and agents, the build vs. buy equation has fundamentally changed. Competitive advantage no longer comes from whether you can build. It comes from rapid iteration on AI that solves real problems for your business and, more importantly, reliably operating those systems at scale. To do that, companies need an internal AI stack as robust as their external product infrastructure. That's exactly what Vercel's agent orchestration platform provides. Build vs. buy ROI has fundamentally changed For decades, the economics of custom internal tools only made sense at large-scale companies. The upfront engineering investment was high, but the real cost was long-term operation with high SLAs and measurable ROI. For everyone else, buying off-the-shelf software w

Introducing new token formats and secret scanning

February 9, 2026

When Vercel API credentials are accidentally committed to public GitHub repositories, gists and npm packages, Vercel now automatically revokes them to protect your account from unauthorized access. When the exposed credentials are detected, you'll receive notifications and can review any discovered tokens and API keys in your dashboard. This detection is powered by GitHub secret scanning and brings an extra layer of security to all Vercel and v0 users. As part of this change, we've also updated token and API key formats to make them visually identifiable. Each credential type now includes a prefix: vcp for Vercel personal access tokens vci for Vercel integration tokens vca for Vercel app access tokens vcr for Vercel app refresh tokens vck for Vercel API keys We recommend reviewing your tokens and API keys regularly, rotating long-lived credentials, and revoking unused ones. Learn more about account security. Read more

How we built AEO tracking for coding agents

February 9, 2026

AI has changed the way that people find information. For businesses, this means it's critical to understand how LLMs search for and summarize their web content. We're building an AI Engine Optimization (AEO) system to track how models discover, interpret, and reference Vercel and our sites. This started as a prototype focused only on standard chat models, but we quickly realized that wasn’t enough. To get a complete picture of visibility, we needed to track coding agents. For standard models, tracking is relatively straightforward. We use AI Gateway to send prompts to dozens of popular models (e.g. GPT, Gemini, and Claude) and analyze their responses, search behavior, and cited sources. Coding agents, however, behave very differently. Many Vercel users interact with AI through their terminal or IDE while actively working on projects. In early sampling, we found that coding agents perform web searches in roughly 20% of prompts. Because these searches happen inline with real development workflows, it’s especially important to evaluate both response quality and source accuracy. Measuring AEO for coding agents requires a different approach than model-only testing. Coding agents aren’t

Introducing Geist Pixel

February 6, 2026

Today, we're expanding the Geist font family with Geist Pixel. Geist Pixel is a bitmap-inspired typeface built on the same foundations as Geist Sans and Geist Mono, reinterpreted through a strict pixel grid. It's precise, intentional, and unapologetically digital. Same system, new texture Geist Pixel isn't a novelty font. It's a system extension. Just like Geist Mono was created for developers, Geist Pixel was designed with real usage in mind, not as a visual gimmick, but as a functional tool within a broader typographic system. It includes five distinct variants, each exported separately: Geist Pixel Square Geist Pixel Grid Geist Pixel Circle Geist Pixel Triangle Geist Pixel Line Every glyph is constructed on a consistent pixel grid, carefully tuned to preserve rhythm, spacing, and legibility. The result feels both nostalgic and contemporary, rooted in early screen typography, but designed for modern products that ship to real users. This matters because pixel fonts often break in production. They don't scale properly across viewports, their metrics conflict with existing typography, or they're purely decorative. Geist Pixel was built to solve these problems, maintaining the visua

Sanity is now available on the Vercel Marketplace

February 6, 2026

Sanity is now available on the Vercel Marketplace as a native CMS integration. Teams can now install, configure, and manage Sanity directly from the Vercel dashboard, eliminating manual API token setup and environment variable configuration. This integration keeps CMS setup inside your existing Vercel workflow instead of requiring a separate dashboard for provisioning and account management. Get started with the integration Define your content schema, set up the client, and start fetching content. Schemas define your content structure in code, specifying document types and their fields. Register your schema types in an index file so Sanity can load them. The Sanity client connects your application to your content. The Marketplace integration provisions the project ID as an environment variable automatically. With the client configured, you can fetch content using GROQ (Graph-Relational Object Queries), Sanity's query language for requesting exactly the fields you need. That's all you need to go from install to fetching content. Install Sanity from the Vercel Marketplace to get started, or deploy the Next.js + Sanity Personal Website template to start from a working example. Read mo

Simplified file retrieval from Vercel Sandbox environments

February 6, 2026

The Vercel Sandbox SDK now includes two new methods that make file retrieval simple. When you run code in a Vercel Sandbox, that code can generate files like a CSV report, a processed image, or a PDF invoice. These files are created inside isolated VMs, so they need to be retrieved across a network boundary. Until now, this required manual stream handling with custom piping. Download a file If you want to download a generated report from your sandbox to your local machine, you can use downloadFile() to seamlessly stream the contents. Read file contents to buffer Both methods handle the underlying stream operations automatically. For example, if your sandbox runs a script that generates a chart as a PNG, you can pull it out with a single call to readFileToBuffer(), no manual stream wiring needed. Learn more about the Sandbox SDK or explore the updated documentation. Read more

The Vercel AI Accelerator is back with $6m in credits

February 5, 2026

Building an AI business is no small feat. Delivering a great agentic product requires infrastructure that handles deployment, security, and scale automatically, but that's table stakes. Startups also need community support, mentorship, investor connections, platform credits, and visibility. That's why we created the Vercel AI Accelerator. Last year, we hosted our second cohort of 40 early-stage teams from across the globe. They joined us for six weeks of learning, building, and shipping, hearing from speakers in leadership at AWS, Anthropic, Cursor, Braintrust, MongoDB, HubSpot, Vercel, and more. The program culminated with a demo day in San Francisco that drew hundreds from across the industry. This year, the Accelerator is back with another cohort of 40 teams building the future of AI. Applications are open now until February 16th. Program benefits The AI Accelerator provides access to thousands of dollars in credits from Vercel, v0, AWS, and a variety of AI platforms. Participants also get to join an exclusive group of AI builders within the Vercel Community. Here are the full details: Credits from Vercel, v0, AWS, and leading AI platforms, including Anthropic, Cursor, ElevenLab

Use Claude Opus 4.6 on AI Gateway

February 5, 2026

Anthropic's latest flagship model, Claude Opus 4.6, is now available on AI Gateway. Built to power agents that handle real-world work, Opus 4.6 excels across the entire development lifecycle. Opus 4.6 is also the first Opus model to support the extended 1M token context window. The model introduces adaptive thinking, a new parameter that lets the model decide when and how much to reason. This approach enables more efficient responses while maintaining quality across programming, analysis, and creative tasks, delivering equal/better performance than extended thinking. Opus 4.6 can interleave thinking and tool calls within a single response. To use the model, set model to anthropic/claude-opus-4.6. The following example also uses adaptive thinking and the effort parameter. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Build logs now support interactive links

February 4, 2026

URLs in build logs are now interactive. Navigate directly to internal and external resources without manually copying and pasting. External links open in a new tab. This eliminates any extra steps you may encounter when investigating build issues or following documentation links. Learn more about accessing build logs. Read more

Parallel's Web Search and tools are live on Vercel

February 4, 2026

You can now use Parallel's LLM-optimized web search and other tools all across Vercel. AI Gateway Unlike provider-specific web search tools that only work with certain models, Parallel's web search tool works universally across all providers. This means you can add web search capabilities to any model without changing your implementation. To use through AI SDK, set parallel_search: gateway.tools.parallelSearch() in tools. Parallel web search extracts relevant excerpts from web pages, making it ideal for agentic tasks and real-time information retrieval. For more control, you can also configure the tool to use specific parameters. For agentic workflows, use mode: 'agentic' to get concise, token-efficient search results that work well in multi-step reasoning. Time-sensitive queries can control cache freshness with maxAgeSeconds, while domain-specific search lets you restrict results to trusted sources or exclude noisy domains. Parallel web search requests are charged at exactly the same rate as the Parallel API. $5 per 1,000 requests (includes up to 10 results per request) and additional results beyond 10 charged at $1 per 1,000 additional results. Read the docs for more information

Parallel joins the Vercel Agent Marketplace

February 4, 2026

Parallel is now available on the Vercel Agent Marketplace with native integration support. Parallel provides web tools and agents designed for LLM-powered applications, including Search, Extract, Tasks, FindAll, and Monitoring capabilities. The Vercel integration provides a single API key that works across all Parallel products, with billing handled directly through your Vercel account. For developers building AI features on Vercel, Parallel enables agents to access the open web for tasks like answering questions, monitoring changes, and extracting structured data. Since Parallel returns results optimized for LLM consumption, your agents can resolve tasks with fewer round trips and reduced cost and latency. Install Parallel from the Marketplace or deploy the Next.js template to see Parallel's web research APIs integrated with Vercel in action. Read more

Turbo build machines by default for new Pro projects

February 3, 2026

Turbo build machines are now the default for all new Pro projects and projects upgrading from Hobby to the Pro plan. Turbo build machines were introduced in October for all paid plans, delivering 30vCPUs and 60GB of memory for faster build performance. Teams adopting Turbo build machines have seen significant build time improvements: up to 30% faster for builds under 2 minutes up to 50% faster for builds that take 2-10 minutes up to 70% faster for builds over 10 minutes Learn more in the documentation or customize your build machine in settings. Read more

Copy visual context to agents from Vercel Toolbar

February 3, 2026

Vercel Toolbar now includes "Copy for Agents" functionality that captures complete visual context from comments, providing coding agents with the technical details they need to understand deployment feedback across your application. When teams copy comments using this feature, agents receive structured context including page URL and viewport dimensions, selected text and node path information, React component tree details, and the original comment text. This helps agents understand exactly where issues occur in your deployed application and what changes are needed. Sample context output: This structured format eliminates the need to manually explain deployment context to agents, enabling direct copying from the toolbar with complete technical details for component location and implementation. The feature is available to all Vercel users immediately. Learn more about Vercel Toolbar or get started with Agents. Read more

Workflow 4.1 Beta: Event-sourced architecture

February 3, 2026

Workflow 4.1 Beta changes how workflows track state internally. Instead of updating records in place, every state change is now stored as an event, and current state is reconstructed by replaying the log. This release also adds support for provider-executed tools and higher throughput. What event sourcing means for workflows Event sourcing is a persistence pattern where state changes are stored as a sequence of events rather than by updating records in place. Instead of storing "this run is completed," the system stores "run_created, then run_started, then run_completed" and reconstructs the current state by replaying those events. In Workflow 4.1, runs, steps, and hooks are no longer mutable database records. They're materializations of an append-only event log. Each event captures a timestamp and context, and the runtime derives current state by processing events in order. This architecture makes workflows more reliable in three ways: Self-healing: If a queue message is lost or a race condition occurs, replaying the workflow route detects missing state and re-enqueues the necessary messages. Old runs required manual intervention to recover from queue downtime; new runs recover au

Zero-configuration support for Koa

February 3, 2026

Vercel now supports Koa applications, an expressive HTTP middleware framework to make web applications and APIs more enjoyable to write, with zero-configuration. Backends on Vercel use Fluid compute with Active CPU pricing by default. This means your Koa app will automatically scale up and down based on traffic, and you only pay for what you use. Visit the Koa on Vercel documentation for more details. Read more

Making agent-friendly pages with content negotiation

February 3, 2026

Agents fetch web pages to answer questions, write code, and complete tasks. When an agent requests a page, it gets everything your browser gets, including navigation menus, stylesheets, JavaScript bundles, tracking scripts, and footer links, when all it needs is the structured text on the page. That extra markup confuses the agent, consumes its context window, and makes every request more expensive. What is content negotiation What agents need is a way to request just the text content of a page, without the browser-specific markup. Content negotiation solves this. It's a standard HTTP mechanism where the client specifies its preferred format via the Accept header, and the server returns the matching representation. Many agents already send Accept: text/markdown when fetching pages, and a server that supports content negotiation can return clean, structured text from the same URL that serves HTML to a browser. We've updated many of our pages, including our blog and changelog, to support content negotiation. This post walks through how it works, how we implemented it in Next.js, and how to add markdown sitemaps so agents can discover your content. How agents request markdown When an

The Vercel OSS Bug Bounty program is now available

February 3, 2026

Security is foundational to everything we build at Vercel. Our open source projects power millions of applications across the web, from small side projects to demanding production workloads at Fortune 500 companies. That responsibility drives us to keep investing in security for the platform and the broader ecosystem. Today, we're opening the Vercel Open Source Software (OSS) bug bounty program to the public on HackerOne. We're inviting security researchers everywhere to find vulnerabilities, challenge assumptions, and help us reduce risk for everyone building with these tools. Since August 2025, we've run a private bug bounty for our open source software with a small group of researchers. That program produced multiple high-severity reports across our Tier 1 projects and helped us refine our processes for triage, fixes, coordinated disclosure, and CVE publication. Now we're ready to expand. Building on our foundation of security investment Last fall, we opened a bug bounty program focused on Web Application Firewall and the React2Shell vulnerability class. Rather than wait for bypasses to surface in the wild, we took a proactive approach: pay security researchers to find them firs

Introducing the new v0

February 3, 2026

Since v0 became generally available in 2024, more than 4 million people have used it to turn their ideas into apps in minutes. v0 has helped people get promotions, win more clients, and work more closely with developers. AI lowered the barrier to writing code. Now we're raising the bar for shipping it. Today, v0 evolves vibe coding from novelty to business critical. Built for production apps and agents, this release includes enterprise-grade security and integrations teams can use to ship real software, not just spin up demos. The limitations of vibe coding We're at an inflection point where anyone can create software. But this freedom has created three problems for the enterprise. Vibe coding is now the world's largest shadow IT problem. AI-enabled software creation is already happening inside every enterprise, and employees are shipping security flaws alongside features: credentials copied into prompts, company data published to the public internet, and databases get deleted, all with no audit trail. Demos are easy to generate, but production features aren't. Prototyping is one of the most popular use cases for marketers and PMs, but the majority of real software work happens on

AI Gateway and one-click deploys now available on TRAE

February 3, 2026

ByteDance's coding agent TRAE now integrates both AI Gateway and direct Vercel deployments, bringing unified AI access and instant production shipping to over 1.6 million monthly active developers. Teams can now access hundreds of models through a single API key and deploy applications directly to Vercel from the TRAE interface. AI Gateway provides unified access to models from Anthropic, OpenAI, Google, xAI, DeepSeek, Z.AI, MiniMax, Moonshot AI, and more without managing multiple provider accounts. The integration includes automatic failover that routes around provider outages, zero markup on AI tokens, and unified observability to monitor both deployments and AI usage. Meanwhile, the Vercel deployment integration handles authorization automatically and returns live URLs immediately after clicking Deploy. SOLO Mode Setting up Vercel deployment In SOLO mode, click the + tab and choose Integrations to connect your Vercel account. When your project is ready, click Deploy in the chat panel to ship directly to production. Once linked, all projects can immediately deploy to Vercel and are also visible in your Vercel dashboard. Setting up AI Gateway In Integrations, choose Vercel AI Gate

Python 3.13 and 3.14 are now available

February 2, 2026

Builds and Functions now support Python 3.13 and Python 3.14 alongside the previously supported Python 3.12. Projects without a specified Python version continue using Python 3.12 by default. The default will switch to Python 3.14 in the coming months. To continue using Python 3.12, specify an upper bound in your project manifest (pyproject.toml or Pipfile) as shown in the examples below. See the Python documentation to learn more about Python support on Vercel. Read more

cubic joins the Vercel Agents Marketplace

January 30, 2026

The Vercel Agents Marketplace now includes cubic, an AI code reviewer that that deploys thousands of AI agents to find and fix bugs in your PRs and codebase. Most code review tools only see what changed. cubic sees how those changes connect to everything else. It learns from your team’s past reviews and gets better over time. Key capabilities include: Catching bugs, regressions, and security vulnerabilities in PRs and existing codebases; continuously running 1000s of agents Identifies senior engineers on your team and learns from their comment history Applying fixes automatically through background agents With cubic handling the first pass, teams spend less time on manual review and more time merging changes. Custom coding standards get enforced across repositories, helping keep code consistent as teams scale. Get started with cubic or explore the Vercel Agents Marketplace to discover more tools. Read more

AssistLoop joins the Vercel Agents Marketplace

January 30, 2026

AssistLoop is now available in the Vercel Marketplace as an AI-powered customer support integration. The integration connects natively with Vercel, so adding AI-driven customer support takes minutes. With AssistLoop, teams can: Install AssistLoop with minimal setup using an Agent ID Add AI-powered support directly to Next.js apps Train agents on internal docs, FAQs, or knowledge bases Customize the assistant to match your brand Review conversations and hand off to human support when needed This integration fits naturally into existing Vercel workflows, with unified billing, automatic environment variables, and no manual configuration. Teams can ship AI-powered support faster without managing separate dashboards or complex setup. AssistLoop automatically injects NEXT_PUBLIC_ASSISTLOOP_AGENT_ID into your project environment. Add the widget script to your site: Get started Deploy the AssistLoop Next.js template from the Marketplace to see it in action. Read more

Run untrusted code with Vercel Sandbox, now generally available

January 30, 2026

AI agents are changing how software gets built. They clone repos, install dependencies, run tests, and iterate in seconds. Despite the change in software, most infrastructure was built for humans, not agents. Traditional compute assumes someone is in the loop, with minutes to provision and configure environments. Agents need secure, isolated environments that start fast, run untrusted code, and disappear when the task is done. Today, Vercel Sandbox is generally available, the execution layer for agents, and we're open-sourcing the Vercel Sandbox CLI and SDK for the community to build on this infrastructure. Built on our compute platform Vercel processes over 2.7 million deployments per day. Each one spins up an isolated microVM, runs user code, and disappears, often in seconds. To do that at scale, we built our own compute platform. Internally code-named Hive, it’s powered by Firecracker and orchestrates microVM clusters across multiple regions. When you click Deploy in v0, import a repo, clone a template, or run vercel in the CLI, Hive is what makes it feel quick. Sandbox brings that same infrastructure to agents. Why agents need different infrastructure Agents don’t work like hum

Vercel Sandboxes are now generally available

January 30, 2026

Vercel Sandboxes are now generally available, providing an ephemeral compute primitive for safely executing untrusted code. It lets teams run AI agent-generated outputs, unverified user uploads, and third-party code without exposing production systems. Each sandbox runs inside Firecracker microVMs, isolated from your infrastructure, so code running in a sandbox is blocked from accessing environment variables, database connections, and cloud resources. Sandboxes are in production use by teams including v0, Blackbox AI and RooCode. To bootstrap a simple Node.js application that creates a Vercel sandbox, use the code below: Or get started with the CLI by opening an interactive shell: Explore the documentation to get started, and check out the open-source SDK and CLI. Read more

Vercel Agent investigations now available in Slack

January 28, 2026

Anomaly alerts proactively monitor your application for usage or error anomalies. When we detect an issue, we send an alert by email, Slack or webhook. Vercel Agent investigates anomaly alerts to find out what's happening in your logs and metrics to help you identify the root cause. With our updated Slack integration, investigations now appear directly in Slack alert messages as a threaded response. This eliminates the need to click into the Vercel dashboard and gives you context to triage the alert directly in Slack. This feature is available for teams using Observability Plus. 10 investigations are included at no additional cost for Observability Plus subscribers. Learn more about Vercel Agent investigations. Read more

Tag-based cache invalidation now available for all responses

January 28, 2026

Vercel's CDN now supports tag-based cache invalidation, giving you granular control over cached content across all frameworks and backends. Responses can now be tagged using the Vercel-Cache-Tag header with a comma-separated list of tags as a new cache organization mechanism to group related content and invalidate it together, rather than just purging your entire cache when content changes. This complements existing headers that cache responses on Vercel's CDN, like Cache-Control, CDN-Cache-Control, and Vercel-CDN-Cache-Control and exposes the same underlying technology that powers Next.js Incremental Static Regeneration (ISR) to any framework or backend. We recommend Next.js applications continue using Incremental Static Regeneration (ISR) for built-in cache tagging and invalidation without managing cache headers manually. How it works After a response has a cache tag, you can invalidate it through dashboard settings, the Vercel CLI, the Function API, or the REST API. Vercel's CDN reads Vercel-Cache-Tag and strips it before sending the response to the client. If you apply cache tags via rewrites from a parent to a child project, and both projects belong to the same team, cached re

How Stripe built a game-changing app in a single flight with v0

January 28, 2026

What would traditionally require months of product-development coordination and building across multiple teams was achieved by one person in a single flight. Read more

Skew Protection now supports prebuilt deployments

January 28, 2026

Skew Protection can now be used with vercel deploy --prebuilt deployments. For teams building locally and uploading with --prebuilt, you can now set a custom deploymentId in your next.config.js: This ID is written to routes-manifest.json and used by Vercel for skew protection routing. You control the ID lifecycle, using the same ID across multiple prebuilt deployments or updating it when deploying new versions. This feature enables Skew Protection support for the specific workflow of building applications locally and then uploading them to Vercel. Learn more about Skew Protection. Read more

How Sensay went from zero to product in six weeks

January 27, 2026

Sensay went from zero to an MVP launch in six weeks by leaning on Vercel previews, feature flags, and instant rollbacks. The team kept one codebase, moved fast through pivots, and shipped without a DevOps team. Read more

AGENTS.md outperforms skills in our agent evals

January 27, 2026

We expected skills to be the solution for teaching coding agents framework-specific knowledge. After building evals focused on Next.js 16 APIs, we found something unexpected. A compressed 8KB docs index embedded directly in AGENTS.md achieved a 100% pass rate, while skills maxed out at 79% even with explicit instructions telling the agent to use them. Without those instructions, skills performed no better than having no documentation at all. Here's what we tried, what we learned, and how you can set this up for your own Next.js projects. The problem we were trying to solve AI coding agents rely on training data that becomes outdated. Next.js 16 introduces APIs like 'use cache', connection(), and forbidden() that aren't in current model training data. When agents don't know these APIs, they generate incorrect code or fall back to older patterns. The reverse can also be true, where you're running an older Next.js version and the model suggests newer APIs that don't exist in your project yet. We wanted to fix this by giving agents access to version-matched documentation. Two approaches for teaching agents framework knowledge Before diving into results, a quick explanation of the two a

Introducing the vercel api CLI command

January 27, 2026

vercel@50.5.1 adds a new api command, giving direct access to the full suite of Vercel APIs from your terminal. The api command provides a direct access point for AI agents to interact with Vercel through the CLI. Agents like Claude Code can access Vercel directly with no additional configuration required. If an agent has access to the environment and the Vercel CLI, it inherits the user's access permissions automatically. List available APIs with vercel api ls, build requests interactively with vercel api, or send requests directly with vercel api [endpoint] [options]. Get started with npx vercel@latest api --help. Read more

Skills v1.1.1: Interactive discovery, open source release, and agent support

January 26, 2026

skills@1.1.1 adds interactive skill discovery and is now fully open source. The new interactive discovery keeps the workflow simple for developers, while also giving agents a clear path to discover skills programmatically by replacing the deprecated npx add-skill command with the updated npx skills interface. You can now use npx skills find to search as you type and discover skills interactively. For AI agents, Skills includes a meta "find-skills" skill, along with a non-interactive mode designed for automated workflows, and support for 27 coding agents. Skills maintenance is also simpler with the new npx skills update command, which refreshes your local skills without manual steps. The full codebase is available on GitHub at Skills. Migration The previous npx add-skill command is deprecated. Use npx skills find for interactive discovery, and use npx skills update to refresh existing skills. Get started with npx skills@latest or explore the Skills repository. Read more

Summaries of CVE-2025-59471 and CVE-2025-59472

January 26, 2026

Two medium-severity denial-of-service vulnerabilities were discovered in self-hosted Next.js applications. Both issues can cause server crashes through memory exhaustion under specific configurations. No data exposure or privilege escalation is possible. Applications hosted on Vercel’s platform are not affected by these issues, and require no customer action. Summary CVE-2025-59471 (CVSS 5.9) affects the Image Optimizer when external image optimization is enabled via remotePatterns. The /_next/image endpoint loads remote images fully into memory without enforcing a maximum size, allowing an attacker to trigger out-of-memory conditions using very large images hosted on an allowed domain. CVE-2025-59472 (CVSS 5.9) affects applications with Partial Pre-Rendering (PPR) enabled in minimal mode. The PPR resume endpoint accepts unauthenticated POST requests and processes attacker-controlled data, allowing memory exhaustion through unbounded request buffering or decompression. Affected Versions CVE-2025-59471 Next.js versions >=10 through =16 through =15 through =16 through <16.1.5 Impact Both vulnerabilities can cause the Node.js process to terminate due to memory exhaustion, resulting in

Summary of CVE-2026-23864

January 26, 2026

Summary Multiple high-severity vulnerabilities in React Server Components were responsibly disclosed. Importantly, these vulnerabilities do not allow for Remote Code Execution. We created new rules to address these vulnerabilities and deployed them to the Vercel WAF to automatically protect all projects hosted on Vercel at no cost. However, do not rely on the WAF for full protection. Immediate upgrades to a patched version are required. Impact React CVE-2026-23864 (CVSS 7.5) CVE-2026-23864 addresses multiple denial of service vulnerabilities in React Server Components. The vulnerabilities are triggered by sending specially crafted HTTP requests to Server Function endpoints, and could lead to server crashes, out-of-memory exceptions or excessive CPU usage; depending on the vulnerable code path being exercised, the application configuration and application code. These vulnerabilities are present in versions 19.0.x, 19.1.x, and 19.2.x of the following packages: react-server-dom-parcel react-server-dom-webpack react-server-dom-turbopack These packages are included in the following frameworks and bundlers: Next.js: 13.x, 14.x, 15.x, and 16.x. Other frameworks and plugins that embed or d

Kimi K2.5 is live on AI Gateway

January 26, 2026

You can now access Kimi K2.5 via AI Gateway with no other provider accounts required. Kimi K2.5 is Moonshot AI's most intelligent and versatile model yet, achieving open-source state-of-the-art performance across agent tasks, coding, visual understanding, and general intelligence. It has more advanced coding abilities compared to previous iterations, especially with frontend code quality and design expressiveness. This enables the creation of fully functional interactive user interfaces with dynamic layouts and animations. To use this model, set model to moonshotai/kimi-k2.5 in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Qwen 3 Max Thinking now available on AI Gateway

January 26, 2026

You can now access Qwen 3 Max Thinking via AI Gateway with no other provider accounts required. Qwen 3 Max Thinking integrates thinking and non-thinking modes for improved performance on complex reasoning tasks. The model autonomously selects and uses its built-in search, memory, and code interpreter tools during conversations without requiring manual tool selection. The tools reduce hallucinations and provide real-time information. To use this model, set model to alibaba/qwen3-max-thinking in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Claude Code Max via AI Gateway, available now for Claude Code

January 26, 2026

AI Gateway now supports the Claude Code Max subscription for the Claude Code CLI. This allows developers to use their existing subscription on Anthropic models with no additional cost while getting unified observability, usage tracking, and monitoring through Vercel’s platform. Setup Set up your environment variables in your shell configuration file (~/.zshrc or ~/.bashrc) Replace your-ai-gateway-api-key with your actual AI Gateway API key. Start Claude Code Log in with your Claude subscription If you're not already logged in, Claude Code will prompt you to authenticate. Choose Option 1 - Claude account with subscription and log in with your Anthropic account. If you encounter issues, try logging out with claude /logout and logging in again. Your Claude Code requests now route through AI Gateway, giving you full visibility into usage patterns and costs while using your Max subscription. How it works When you configure Claude Code to use AI Gateway, Claude Code continues to authenticate with Anthropic. It sends its Authorization header and AI Gateway acts as either a passthrough proxy to Anthropic or, when it needs to fall back, a router to other providers. Since the Authorization h

Live model performance metrics accessible via AI Gateway

January 26, 2026

AI Gateway now displays throughput and latency metrics across hundreds of models, helping you choose the right model based on live performance data. Metrics appear in three places and are updated every hour: Model list: Best performance per model (P50 latency and throughput) Model detail pages: Provider-level performance breakdown REST API: Rolling endpoint performance aggregates (latency and throughput, P50/P95) Model list The AI Gateway model list now includes sortable columns for latency and throughput. Each row displays the best P50 metrics (lowest latency, highest throughput) for that model across all its available providers. Metrics are updated every hour and based on live AI Gateway customer requests. Sort by throughput to find the fastest token generation, or by latency to find models with the quickest time-to-first-token. Model detail pages On the individual model pages, you can see P50 latency and throughput for each provider that has recorded usage. This helps you compare provider performance for the same model and choose the best option for your use case. To access these pages, click on any model in the model list to get a more detailed view of the breakdown across all

Agent skills explained: An FAQ

January 26, 2026

Learn what agents skills are, how to install them, how agents use them, and best practices for implementation. Read more

Trinity Large Preview is on AI Gateway

January 26, 2026

You can now access Trinity Large Preview via AI Gateway with no other provider accounts required. Trinity Large Preview is optimized for reasoning-intensive workloads, including math, coding tasks, and complex multi-step agent workflows. It is designed to handle extended multi-turn interactions efficiently while maintaining high inference throughput. To use this model, set model to arcee-ai/trinity-large-preview in the AI SDK: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Use AI Gateway with Clawdbot

January 24, 2026

Clawdbot is a personal AI assistant powered by Claude with persistent memory. It can browse the web, run shell commands, and manage files across any operating system. You can use Clawdbot with Vercel AI Gateway to access hundreds of models from multiple providers through a single endpoint. AI Gateway provides unified API access across models without managing separate API keys. Create an API key in the AI Gateway dashboard, then install Clawdbot: Run the onboarding wizard: Select Vercel AI Gateway as your provider and enter your AI Gateway API key. You can then choose from hundreds of available models. Your AI assistant is now running and ready to help with tasks across your system. See the AI Gateway docs for more details on Clawdbot and more integrations. Read more

Vercel now supports customizing platform error pages

January 23, 2026

You can now customize error pages for platform errors on Vercel, replacing generic error pages with your own branded experiences. Custom error pages display when Vercel encounters uncaught errors like function invocation timeouts or other platform errors. How it works You can implement custom error pages using your framework’s conventions and Vercel will automatically locate them, for example with Next.js you can simply place a 500/page.tsx or static 500.html page in the public directory. To enrich error pages with request-specific context, you can use the following metadata tokens: ::vercel:REQUEST_ID:: - Contains the Vercel request ID ::vercel:ERROR_CODE:: - The specific error code e.g. FUNCTION_INVOCATION_TIMEOUT This feature is available for Enterprise teams and enabled automatically across all projects. No additional configuration required. See the documentation to get started or reference the following implementations: Custom error pages with App Router or Custom error pages with public directory. Read more

Configure build machine settings across all projects

January 23, 2026

Build and deployment settings can now be configured at the team level and applied across all projects, compared to the previous project-by-project setup. Build Machines let you choose the compute resources for each build to optimize build times: Standard build machines with 4 vCPUs and 8 GB of memory Enhanced build machines with 8 vCPUs and 16 GB of memory Turbo build machines with 30 vCPUs and 60 GB of memory On-Demand Concurrent Builds control how many builds can run in parallel and whether builds skip the queue. You can now apply configurations to all projects at once, or make targeted changes across multiple projects from a single interface. Get started with team-level settings. Read more

Faster deploys with improved function caching

January 23, 2026

Function uploads are now skipped when code hasn't changed, reducing build times by 400-600ms on average and up to 5 seconds for larger builds. Previously, deployment-specific environment variables like VERCEL_DEPLOYMENT_ID were included in the function payload, making every deployment unique even with identical code. These variables are now injected at runtime, allowing Vercel to recognize unchanged functions and skip redundant uploads. This optimization applies to Vercel Functions without a framework, and projects using Python, Go, Ruby, and Rust. Next.js projects will receive the same improvement soon. The optimization is applied automatically to all deployments with no configuration required. Learn more about functions and builds in our documentation. Read more

Testing if "bash is all you need"

January 22, 2026

We invited Ankur Goyal from Braintrust to share how they tested the "bash is all you need" hypothesis for AI agents. There's a growing conviction in the AI community that filesystems and bash are the optimal abstraction for AI agents. The logic makes sense: LLMs have been extensively trained on code, terminals, and file navigation, so you should be able to give your agent a shell and let it work. Even non-coding agents may benefit from this approach. Vercel's recent post on building agents with filesystems and bash showed this by mapping sales calls, support tickets, and other structured data onto the filesystem. The agent greps for relevant sections, pulls what it needs, and builds context on demand. But there's an alternative view worth testing. Filesystems may be the right abstraction for exploring and retrieving context, but what about querying structured data? We built an eval harness to find out. Setting up the eval We tasked agents with querying a dataset of GitHub issues and pull requests. This type of semi-structured data mirrors real-world use cases like customer support tickets or sales call transcripts. Question complexity ranged from: Simple queries: "How many open iss

New dashboard navigation available

January 22, 2026

A redesign of the navigation in the dashboard is now available as an opt-in experience. This new navigation maintains full functionality while streamlining access to your most-used features. New Sidebar — Moved horizontal tabs to a resizable sidebar that can be hidden when not needed Consistent Tabs — Unified sidebar navigation with consistent links across team and project levels Improved Order — Reordered navigation items to prioritize the most common developer workflows Projects as Filters — Switch between team and project versions of the same page in one click Optimized for Mobile — New mobile navigation featuring a floating bottom bar optimized for one-handed use Try the new navigation today before it rolls out to all users. Read more

Filesystem snapshots supported on Vercel Sandboxes

January 22, 2026

Vercel Sandbox now supports filesystem snapshots to capture your state. You can capture a Sandbox's complete filesystem state as a snapshot and launch new Sandboxes from that snapshot using the Sandbox API. This eliminates repeated setup when working with expensive operations like dependency installation, builds, or fixture creation. Create the environment once, snapshot it, then reuse that exact filesystem state across multiple isolated runs. How snapshots work Snapshots capture the entire filesystem of a running Sandbox. New Sandboxes can launch from that snapshot, providing immediate access to pre-installed dependencies and configured environments. Key capabilities Create a snapshot from any running Sandbox with sandbox.snapshot() Launch new Sandboxes from snapshots via source: { type: 'snapshot', snapshotId } Reuse the same snapshot with multiple Sandboxes for parallel testing and experimentation See the documentation to get started with snapshots. Read more

Apply code suggestions from Vercel Agent with one click

January 21, 2026

You can now apply suggested code fixes from the Vercel Agent directly in the Vercel Dashboard. When the Vercel Agent reviews your pull request, suggestions include a View suggestion button that lets you commit the fix to your PR branch, including changes that touch multiple files. Suggestions open in the dashboard, where you can accept them in bulk or apply them one by one. After you apply a suggestion, the review thread is automatically resolved. You can also track multiple concurrent Vercel Agent jobs from the Tasks page. Get started with Vercel Agent code review in the Agent dashboard, or learn more in the documentation. Read more

Cron Jobs now visible in Deployment Summary

January 21, 2026

You can now view the Cron Jobs from your application in the Deployment Summary section of Deployments. Try it out by deploying a Vercel Cron Job template. Once you deploy, Vercel automatically registers your cron jobs. Learn more in the Cron Jobs documentation. Read more

AI Code Elements

January 21, 2026

Today we're releasing a brand new set of components designed to help you build the next generation of IDEs, coding apps and background agents. A composable component for displaying an AI SDK ToolLoopAgent configuration with model, instructions, tools, and output schema. Building on what we've learned from Streamdown, we massively improved the code block component with support for a header, icon, filename, multiple languages and a more performant renderer. The Commit component displays commit details including hash, message, author, timestamp, and changed files. The EnvironmentVariables component displays environment variables with value masking, visibility toggle, and copy functionality. The FileTree component displays a hierarchical file system structure with expandable folders and file selection. The PackageInfo component displays package dependency information including version changes and change type badges. The Sandbox component provides a structured way to display AI-generated code alongside its execution output in chat conversations. It features a collapsible container with status indicators and tabbed navigation between code and output views. The SchemaDisplay component vis

Use skills in your AI SDK agents via bash-tool

January 21, 2026

Skills support is now available in bash-tool, so your AI SDK agents can use the skills pattern with filesystem context, Bash execution, and sandboxed runtime access. This gives your agent a consistent way to pull in the right context for a task, using the same isolated execution model that powers filesystem-based context retrieval. This allows giving your agent access to the wide variety of publicly available skills, or for you to write your own proprietary skills and privately use them in your agent. Read the bash-tool changelog for background and check out createSkillTool documentation. Read more

Introducing the Montréal, Canada region (yul1)

January 20, 2026

Montréal, Canada (yul1) is now part of Vercel’s global delivery network, expanding our footprint to deliver lower latency and improved performance for users in Central Canada. The new Montréal region extends our globally distributed CDN’s caching and compute closer to end users, reducing latency without any changes required from developers. Montréal is generally available and handling production traffic. Teams can configure Montréal as an execution region for Vercel Functions, powered by Fluid compute to enhance resource efficiency, minimize cold starts, and scale automatically with demand. Teams with Canadian data residency requirements can also use Montréal to keep execution in Canada. Learn more about Vercel Regions and Montréal regional pricing Read more

Introducing skills, the open agent skills ecosystem

January 20, 2026

We released skills, a CLI for installing and managing skill packages for agents. Install a skill package with npx skills add . So far, skills has been used to install skills on: amp, antigravity, claude-code, clawdbot, codex, cursor, droid, gemini, gemini-cli, github-copilot, goose, kilo, kiro-cli, opencode, roo, trae, and windsurf. Today we’re also introducing skills.sh, a directory and leaderboard for skill packages. Use it to: discover new skills to enhance your agents browse skills by category and popularity track usage stats and installs across the ecosystem Get started with npx skills add vercel-labs/agent-skills and explore skills.sh. Read more

Cron jobs now support 100 per project on every plan

January 20, 2026

Cron jobs on Vercel no longer have per-team limits, and per-project limits were lifted to 100 on all plans. Previously, all plans had a cap of 20 cron jobs per project, with per-team limits of 2 for Hobby, 40 for Pro, and 100 for Enterprise. To get started, add cron entries to vercel.json: You can also deploy the Vercel Cron Job template. Once you deploy, Vercel automatically registers your cron jobs. Learn more in the Cron Jobs documentation. Read more

Recraft image models now on AI Gateway

January 19, 2026

Recraft models are now available via Vercel's AI Gateway with no other provider accounts required. You can access Recraft's image models, V3 and V2. These image models excel at photorealism, accurate text rendering, and complex prompt following. V3 supports long multi-word text generation with precise positioning, anatomical correctness, and native vector output. It includes 20+ specialized styles from realistic portraits to pixel art. To use this model, set model to recraft/recraft-v3 in the AI SDK. This model supports generateImage. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Improved environment variables UI

January 16, 2026

The environment variables UI is now easier to manage across shared and project environment variables. You can spend less time scrolling, use larger hit targets, and view details only when you need them. Learn more in the environment variables documentation. Read more

SSH into running Vercel Sandboxes with the CLI

January 15, 2026

You can now open secure, interactive shell sessions to running Sandboxes with the Vercel Sandbox CLI. Note: While you’re connected, the Sandbox timeout is automatically extended in 5-minute increments to help avoid unexpected disconnections, for up to 5 hours. Learn more in the Sandbox CLI docs. Read more

Experimental build mode for Hono and Express projects

January 15, 2026

Users can opt in to an experimental build mode for Hono and Express projects, which lets you filter logs by route, similar to Next.js. It also updates the build pipeline with better module resolution: Relative imports no longer require file extensions TypeScript path aliases are supported Improved ESM and CommonJS interoperability To enable it, set VERCEL_EXPERIMENTAL_BACKENDS=1 in your project's environment variables. Read more

OpenResponses API now supported on Vercel AI Gateway

January 15, 2026

Vercel AI Gateway is a day 0 launch partner for the OpenResponses API, an open-source specification from OpenAI for multi-provider AI interactions. OpenResponses provides a unified interface for text generation, streaming, tool calling, image input, and reasoning across providers. AI Gateway supports OpenResponses for: Text generation: Send messages and receive responses from any supported model. Streaming: Receive tokens as they're generated via server-sent events. Tool calling: Define functions that models can invoke with structured arguments. Image input: Send images alongside text for vision-capable models. Reasoning: Enable extended thinking with configurable effort levels. Provider fallbacks: Configure automatic fallback chains across models and providers. Use OpenResponses with your AI Gateway key, and switch models across providers by changing the model string. You can also use OpenResponses for more complex cases, like tool calling. Read the OpenResponses API documentation or view the specification. Read more

On-demand Vercel Agent code reviews

January 15, 2026

You can now trigger a Vercel Agent code review on demand. When Vercel post comments on your GitHub pull request, you can now click the Review with Vercel Agent button from the deployment table to trigger a code review. If you want more control, enable automatic code reviews in Team Settings → Agent → Review PRs Automatically. Get started by installing the Vercel GitHub App or read the documentation. Read more

AWS databases are now live on the Vercel Marketplace and v0

January 15, 2026

AWS databases are now available in the Vercel Marketplace and v0. Read more

Access Perplexity Web Search on Vercel AI Gateway with any model

January 14, 2026

You can now give any model the ability to search the web using Perplexity through Vercel's AI Gateway. AI Gateway supports Perplexity Search as a universal web search tool that works with all models, regardless of provider. Unlike native search tools that are exclusive to specific providers, Perplexity Search can be added to all models. To use Perplexity Search with the AI SDK, import gateway.tools.perplexitySearch() from @ai-sdk/gateway and pass it in the tools parameter as perplexity_search to any model. Some example use cases include: • Models without native search: Enable web search on models like zai/glm-4.7 or any from any other providers that don't expose a built-in search tool. • Developer tooling and CI assistants: Get current package versions, recently merged PRs, release notes, or docs updates. • Consistency with fallbacks: Maintain search behavior across multiple providers without rewriting search logic. For more information, see the AI Gateway Perplexity Web Search docs. Read more

GPT 5.2 Codex now available on Vercel AI Gateway

January 14, 2026

You can now access GPT 5.2 Codex with Vercel's AI Gateway and no other provider accounts required. GPT 5.2 Codex combines GPT 5.2's strength in professional knowledge work with GPT 5.1 Codex Max's agentic coding capabilities. GPT 5.2 Codex is better at working on long running coding tasks compared to predecessors and can handle more complex tasks like large refactors and migrations more reliably. The model has stronger vision performance for more accurate processing of screenshots and charts that are shared while coding. GPT 5.2 Codex also surpasses GPT 5.1 Codex Max in cyber capabilities and outperformed the previous model in OpenAI's Professional Capture-the-Flag (CTF) cybersecurity eval. To use the GPT 5.2 Codex, with the AI SDK, set the model to openai/gpt-5.2-codex: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Reduced build times for large projects

January 14, 2026

We shipped build system optimizations that reduce overhead for projects with many input files, large node_modules, or large build outputs. Expensive disk operations (large file detection and folder size calculations) are no longer on the critical path for successful builds. These calculations now only run when a build fails, or when you enable the VERCEL_BUILD_SYSTEM_REPORT environment variable. Builds complete 2.8 seconds faster on average, with larger builds seeing improvements of up to 12 seconds. See the builds documentation for details. Read more

Use Perplexity Web Search with Vercel AI Gateway

January 14, 2026

Models are powerful, but they're limited to their training data and knowledge cutoff date. When users ask about today's news, current prices, or the latest API changes, models can offer outdated information or admit they don't know. Provider-agnostic web search on AI Gateway changes this. With a single line of code, you can give any model the ability to search the web in real-time. It works with OpenAI, Anthropic, Google, and every other provider available through AI Gateway. Read more

Introducing: React Best Practices

January 14, 2026

We've encapsulated 10+ years of React and Next.js optimization knowledge into react-best-practices, a structured repository optimized for AI agents and LLMs. Read more

Node.js runtime now defaults to version 24 for Vercel Sandbox

January 14, 2026

Vercel Sandbox for Node.js now uses Node.js 24 by default. This keeps the Node.js runtime aligned with the latest Node.js features and performance improvements. If you don’t explicitly configure a runtime, Sandbox will use Node.js 24 (as shown below). Read the Sandbox documentation to learn more. Read more

Docs pages support Markdown responses

January 13, 2026

You can now request Vercel documentation as Markdown by sending the Accept header with the value text/markdown. This makes it easier to use docs content in agentic and CLI workflows, indexing pipelines, and tooling that expects text/markdown. Markdown responses include a sitemap.md link at the end. You and your agent can use it to discover additional docs pages programmatically. Read more

Project-level deployment suffixes

January 13, 2026

You can now set deployment suffixes at the project level, instead of using a single suffix across your team. Project-level suffixes follow the same requirements as team-level suffixes. Your team must own the domain, and the domain must use Vercel nameservers. Configure suffixes in your project settings. Read more

Set team-wide defaults for Deployment Protection

January 13, 2026

Set team-wide defaults for Deployment Protection You can now set a team-wide default for Deployment Protection. New projects start with Deployment Protection set to Standard Protection, which protects Preview Deployments by default. Choose the default protection level for new projects: All Deployments, Standard Protection, or None. If you set All Deployments as the default, every new project is protected as soon as it’s created. See the Deployment Protection documentation. Read more

Protection bypass for automation now supports multiple secrets

January 13, 2026

Vercel projects now support multiple Protection Bypass for Automation secrets. This makes it easier to rotate secrets and use different secrets for different workflows. Each bypass can also include a note, so it’s easier to track what it’s used for. Learn more in the Deployment Protection documentation. Read more

Nick Bogaty joins Vercel as Chief Revenue Officer

January 13, 2026

It's a thrilling time to work in Sales at Vercel. The web is transitioning from pages to agents, and Vercel is building the self-driving infrastructure to power it. We've assembled a Sales organization that equally understands the continually shifting technical landscape and pressing business needs to stay flexible, move fast, and be secure in the AI era. We're rethinking how Sales operates, and we're building the most AI-forward go-to-market organization in the industry. To lead this charge, we're welcoming Nick Bogaty as our Chief Revenue Officer. Read more

Build cache storage increased for larger build machines

January 12, 2026

Projects that use larger build machines now have increased limits for build caches. Enhanced build machines have a 3 GB limit Turbo build machines have a 4.75 GB limit These limit increases come at no additional charge. Learn more about Vercel's build cache. Read more

Streamdown v2: Smaller bundle size and new Remend options

January 12, 2026

Today, we're releasing a major update to Streamdown, our drop-in replacement for react-markdown, designed for AI-powered streaming. Streamdown Plugins The most requested feature since launch has been to reduce the bundle size. Streamdown v2 now ships with a bundle size much smaller than the previous version and uses a plugin-based architecture. Carets Streamdown now includes built-in caret (cursor) indicators that display at the end of streaming content. Carets provide a visual cue to users that content is actively being generated, similar to a blinking cursor in a text editor. Configurable Remend Our underlying markdown-healing library Remend is now configurable. You can choose how much healing you would prefer during the markdown streaming. Read more

How Mux shipped durable video workflows with their @mux/ai SDK

January 12, 2026

We invited Dylan Jhaveri from Mux to share how they shipped durable workflows with their @mux/ai SDK. AI workflows have a frustrating habit of failing halfway through. Your content moderation check passes, you're generating video chapters, and then you hit a network timeout, a rate limit, or a random 500 from a provider having a bad day. Now you're stuck. Do you restart from scratch and pay for that moderation check again? Or do you write a bunch of state management code to remember where you left off? This is where durable execution changes everything. When we set out to build @mux/ai, an open-source SDK to help our customers build AI features on top of Mux's video infrastructure, we faced a fundamental question: how do we ship durable workflows in a way that's easy for developers to adopt, without forcing them into complex infrastructure decisions? The answer was Vercel's Workflow DevKit. Read more

Web Interface Guidelines now available as an agent command

January 12, 2026

You can now install Vercel's Web Interface Guidelines as skill/command for your agent. Run /web-interface-guidelines to review UI code for accessibility, keyboard support, form behavior, animation, performance, and more. Install with a single command: curl -fsSL https://vercel.com/design/guidelines/install | bash Supports Claude Code, Cursor, OpenCode, Windsurf, and Gemini CLI. For other agents, use the command prompt directly or add AGENTS.md to your project. Learn more about using the Web Interface Guidelines with Agents. Read more

How to build agents with filesystems and bash

January 9, 2026

Many of us have built complex tooling to feed our agents the right information. It's brittle because we're guessing what the model needs instead of letting it find what it needs. We've found a simpler approach. We replaced most of the custom tooling in our internal agents with a filesystem tool and a bash tool. Our sales call summarization agent went from ~news: [.00 to ~$0.25 per call on Claude Opus 4.5, and the output quality improved. We used the same approach for d0, our text-to-SQL agent. The idea behind this is that LLMs have been trained on massive amounts of code. They've spent countless hours navigating directories, grepping through files, and managing state across complex codebases. If agents excel at filesystem operations for code, they'll excel at filesystem operations for anything. Agents already understand filesystems. Customer support tickets, sales call transcripts, CRM data, conversation history. Structure it as files, give the agent bash, and the model brings the same capabilities it uses for code navigation. Read more

Limit on-demand concurrent builds to one build per branch

January 9, 2026

On-Demand Concurrent Builds let builds start immediately, without waiting for other deployments to finish. You can now configure this feature to run one active build per branch. When enabled, deployments to the same branch are queued. After the active build finishes, only the most recent queued deployment starts building. Older queued deployments are skipped. Deployments on different branches can still build concurrently. Enable this in your project settings or learn more in the documentation. Read more

Bookmark domains on Vercel Domains

January 8, 2026

You can now bookmark domains on Vercel Domains for purchasing at a later date. To save a domain, either: Click on a search result, and select "Save for later" Select the bookmark icon on a domain in your cart You can then view your saved domains and add them to your cart from the "Saved" tab. Try it now at vercel.com/domains Read more

Secure Compute is now self-serve

January 7, 2026

Teams can now create, update and delete Secure Compute networks directly from the Vercel dashboard, the API, and Terraform. Secure Compute networks provide private connectivity between your Vercel Functions and backend infrastructure and let you control regional placement, addressing, egress and failover of your projects. Now you can: Self-service network management with no contract amendment or manual provisioning required. ‍ Manage existing Secure Compute capabilities directly, including Region and Availability Zone selection, active/passive failover, private CIDR selection, NAT/egress behavior are now manageable via self-serve flows. Automate & integrate with full network lifecycle support through the Dashboard, public API, and Terraform so teams can manage networks interactively or declaratively. And coming soon: Self-serve Site-to-Site VPN connections via the Dashboard, API, and Terraform, Secure Compute for Pro customers and PrivateLink connectivity. This is available today for Enterprise teams. Check out the documentation to get started.​​​​‌‍​‍​‍‌‍ Read more

How we made v0 an effective coding agent

January 7, 2026

Last year we introduced the v0 Composite Model Family, and described how the v0 models operate inside a multi-step agentic pipeline. Three parts of that pipeline have had the greatest impact on reliability. These are the dynamic system prompt, a streaming manipulation layer that we call “LLM Suspense”, and a set of deterministic and model-driven autofixers that run after (or while!) the model finishes streaming its response. What we optimize for The primary metric we optimize for is the percentage of successful generations. A successful generation is one that produces a working website in v0’s preview instead of an error or blank screen. But the problem is that LLMs running in isolation encounter various issues when generating code at scale. In our experience, code generated by LLMs can have errors as often as 10% of the time. Our composite pipeline is able to detect and fix many of these errors in real time as the LLM streams the output. This can lead to a double-digit increase in success rates. Read more

Introducing bash-tool for filesystem-based context retrieval

January 7, 2026

We open-sourced bash-tool, the Bash execution engine used by our text-to-SQL agent that we recently re-architected to reduce our token usage, improve the accuracy of the agent's responses, and improve the agent's overall performance. bash-tool gives your agent a way to find the right context by running bash-like commands over files, then returning only the results of those tool calls to the model. Context windows can fill up quickly if you include large amounts of text into a prompt. As agents tend to do well with Unix-style workflows like find, grep, jq, and pipes, with bash-tool you can now keep large context local, in a filesystem, and let the agent use those commands to retrieve smaller slices of context on demand. bash-tool provides bash, readFile, and writeFile tools for AI SDK agents, working with both in-memory and sandboxed environments, and: runs on top of just-bash, which interprets bash scripts directly in TypeScript without a shell process or arbitrary binary execution you can preload that filesystem with your files at startup, so your agent can search them when needed without pasting everything into the prompt it supports running in-memory or in a custom isolated VM I

Vercel Agent code reviews now follow your code guidelines

January 6, 2026

Vercel Agent now applies your repository’s coding guidelines during code reviews. Add an AGENTS.md file to your repository, or use existing formats like CLAUDE.md, .cursorrules, or .github/copilot-instructions.md. Agent automatically detects and applies these guidelines to provide context-specific feedback for your codebase. No configuration required. Learn more about code guidelines. Read more

AI Gateway support for Claude Code

January 5, 2026

You can now use Claude Code through Vercel AI Gateway via its Anthropic-compatible API endpoint. Route Claude Code requests through AI Gateway to centralize usage and spend, view traces in observability, and benefit from failover between providers for your model of choice. Log out if you're already logged in, then set these environment variables to configure Claude Code to use AI Gateway: Setting ANTHROPIC_API_KEY to an empty string is required. Claude Code checks this variable first, and if it's set to a non-empty value, it will use that instead of ANTHROPIC_AUTH_TOKEN. Start Claude Code. Requests will route through AI Gateway: See the Claude Code documentation for details. Read more

Stopping the slow death of internal tools

December 27, 2025

Companies spend millions of dollars in time and money trying to build internal tools. These range from lightweight automations and dashboards to fully custom systems with dedicated engineering teams. Most businesses can’t justify focusing developers on bespoke operational tools, so non-technical teams resort to brittle and insecure workarounds: custom Salesforce formulas and fields, complex workflow automations, spreadsheets, and spiderwebs of integrations across platforms. They are trying to build software without actually building software, and most of the tools end up collecting dust. v0’s AI agent changes this equation. Business users can build and publish real code and apps on the same platform that their developers use, safely integrate with internal and external systems, and secure everything behind existing SSO authentication. Read more

Pixel Portraits: AI generated trading cards

December 23, 2025

At our recent Next.js Conf and Ship AI events, we introduced an activation that blended technical experimentation with playful nostalgia. The idea started long before anyone stepped into the venue. As part of the online registration experience for both events, attendees could prompt and generate their own trading cards, giving them an early taste of the format and creating the foundation for what we wanted to bring into the real world. Read more

MiniMax M2.1 now live on Vercel AI Gateway

December 22, 2025

You can now access MiniMax's latest model, M2.1, with Vercel's AI Gateway and no other provider accounts required. MiniMax M2.1 is faster than its predecessor M2, with clear improvements specifically in coding use cases and complicated multi-step tasks with tool calls. M2.1 writes higher quality code, is better at following instructions for difficult tasks, and has a cleaner reasoning process. The model has breadth in addition to depth, with improved performance across multiple coding languages (Go, C++, JS, C#, TS, etc.) and refactoring, feature adds, bug fixes, and code review. To start building with MiniMax M2.1 via AI SDK, set the model to minimax/minimax-m2.1: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

GLM-4.7 available on Vercel AI Gateway

December 22, 2025

You can now access Z.ai's latest model, GLM-4.7, with Vercel's AI Gateway and no other provider accounts required. GLM-4.7 comes with major improvements in coding, tool usage, and multi-step reasoning, especially with complex agentic tasks. The model also has a more natural tone for a better conversational experience and can product a more refined aesthetic for front-end work. To start building with GLM-4.7 via AI SDK, set the model to zai/glm-4.7: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Function start type now available in Runtime Logs

December 22, 2025

For any request involving a Vercel Function invocation, you can now view the function start type in the right hand details panel of Runtime Logs. A Function invocation can be either: Hot, Hot (prewarmed) or Cold. When a Function was invoked and it's a Cold start, we also display the cold start duration like: Cold (280ms). Try it out or learn more about Runtime Logs. Read more

We removed 80% of our agent’s tools

December 22, 2025

It got better. We spent months building a sophisticated internal text-to-SQL agent, d0, with specialized tools, heavy prompt engineering, and careful context management. It worked… kind of. But it was fragile, slow, and required constant maintenance. So we tried something different. We deleted most of it and stripped the agent down to a single tool: execute arbitrary bash commands. We call this a file system agent. Claude gets direct access to your files and figures things out using grep, cat, and ls. The agent got simpler and better at the same time. 100% success rate instead of 80%. Fewer steps, fewer tokens, faster responses. All by doing less. Read more

AI SDK 6

December 22, 2025

With over 20 million monthly downloads and adoption by teams ranging from startups to Fortune 500 companies, the AI SDK is the leading TypeScript toolkit for building AI applications. It provides a unified API, allowing you to integrate with any AI provider, and seamlessly integrates with Next.js, React, Svelte, Vue, and Node.js. The AI SDK enables you to build everything from chatbots to complex background agents. Read more

Introducing vercel.ts: Programmatic project configuration

December 19, 2025

Vercel now supports vercel.ts, a new TypeScript-based configuration file that brings type safety, dynamic logic, and better developer experience to project configuration. vercel.ts lets you express configuration as code by defining advanced routing, request transforms, caching rules, and cron jobs, going beyond what static JSON can express. In addition to full type safety, this also allows access to environment variables, shared logic, and conditional behavior. All projects can now use vercel.ts (or .js, .mjs, .cjs, .mts) for project configuration. Properties are defined identically to vercel.json and can be enhanced using the new @vercel/config package. Try the playground to explore vercel.ts, learn how to migrate from an existing vercel.json, or read the documentation and the @vercel/config package. Read more

Bulk redirects UI, API, and CLI now generally available

December 18, 2025

Vercel users can now configure bulk redirects using UI, API, or CLI without a new deployment. Vercel's bulk redirects allow up to one million static URL redirects per project. In addition to bulk redirects support via vercel.json, these new changes simplify how teams can manage large-scale migrations, quickly fix broken links, handle expired pages, and more. You can modify redirects individually, or in bulk by uploading CSV files. Redirect changes are initially staged for testing before publishing to production, and include a version history to see and restore historical versions. This feature is available for Pro and Enterprise customers, with rates for additional capacity: Pro: 1,000 bulk redirects included per project Enterprise: 10,000 bulk redirects included per project Additional capacity: starts at $50/month per 25,000 redirects Get started with bulk redirects, or learn more. Read more

Chat with Vercel Marketplace integrations using Vercel Agent

December 18, 2025

You can now interact with installed Marketplace integrations using Vercel Agent in the Dashboard. This feature launches with support from Marketplace providers including Neon, Supabase, Dash0, Stripe, Prisma and Mux, with more coming soon. You can use Vercel Agent, a chat-based interface to talk to Marketplace Providers MCP (Model Context Protocol), allowing you to query, debug, and manage connected services directly from Vercel. Tools exposed by providers are available automatically, with authentication and configuration handled by Vercel. Available free for Vercel Pro and Enterprise customers, with an optional Read-Only mode for safe exploration and debugging. How to get started Install or visit a supported Marketplace integration Click on Agent Tools in the left navigation to open the chat interface. Your installed integration's tools load automatically and are ready to use. Learn more and get started in the documentation. Read more

Reduced prices for TLDs .site, .space, .website, .fun, .online, .store, .tech

December 18, 2025

Vercel Domains now offers reduced prices for the following TLDs: .site: Now news: [.99, down from $2.99 .space: Now news: [.99, down from $4.99 .website: Now news: [.99, down from $4.99 .fun: Now news: [.99, down from $4.99 .online: Now news: [.99, down from $2.99 .store: Now news: [.99, down from $2.99 .tech: Now $7.99, down from news: [3.99 Prices for premium domains are not affected by this pricing change. Get your domain today at vercel.com/domains. Read more

Preview URLs optimized for multi-tenant platforms

December 17, 2025

Vercel helps you create multi-tenant platforms, where a single project can be backed by tens of thousands of domains, like vibe coding platforms, website builders, e-commerce storefronts and more. We're making it even easier to build those styles of apps today by introducing dynamic URL prefixes. Dynamic URL prefixes allow you to prefix your existing deployment urls with {data}---, for example tenant-123---project-name-git-branch.yourdomain.dev This will route the traffic to project-name-git-branch.yourdomain.dev while keeping tenant-123--- in the url which your app can extract and route based on it. Previously, preview URLs were designed to match a specific preview deployment exactly and Vercel wouldn’t have enough information to route domains to a specific preview deployment. Now you can: Create unique preview URLs for each tenant Encode metadata, routing context, or automation signals directly in the URL Use flexible URL structures such as: tenant-123---project-name-git-branch.yourdomain.dev Preview URLs for multi-tenant platforms are available for Pro and Enterprise teams, and require a Preview Deployment Suffix (a Pro add-on). Try the demo or to get started, go to your team's

AWS databases now available on the Vercel Marketplace

December 17, 2025

Today we’re introducing native support for AWS databases including Amazon Aurora PostgreSQL, Amazon Aurora DSQL, and Amazon DynamoDB on the Vercel Marketplace. This gives developers a direct path to provision and manage scalable, production-ready AWS databases from within the Vercel dashboard with no manual setup required, and: One-click support for creating a new AWS account, provisioning new AWS databases and linking them to your Vercel projects. Improved developer experience with simplified region selection, secure credential handling, and unified monitoring of AWS database resources from Vercel. Automatic environment variable for connection strings and credentials, securely stored within your Vercel project. Free starter plan for new AWS customers, including news: [00 in credits, with deep links to manage or upgrade plans in the AWS console. And coming soon: Provision databases into your existing AWS account, attach them to your projects, and access AWS databases directly inside v0. Getting started Navigate to the Vercel Marketplace and select AWS Choose Create new account to provision a database Select your database type, region, and plan (including a free starter plan with news: [00 in

Gemini 3 Flash is now available on the Vercel AI Gateway

December 17, 2025

You can now access Google's latest Gemini model, Gemini 3 Flash, with Vercel's AI Gateway and no other provider accounts required. It is Google's most intelligent model that is optimized for speed, with Gemini 3's pro-grade reasoning alongside flash-level latency, efficiency, and cost. Gemini 3 Flash significantly outperforms the previous Gemini 2.5 models, beating Gemini 2.5 Pro across most benchmarks, while using 30% less tokens and is 3x faster at a fraction of the cost. To use the Gemini 3 Flash with the AI SDK, set the model to google/gemini-3-flash: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Cline now runs on Vercel AI Gateway

December 16, 2025

Cline, the leading open-source coding agent built for developers and teams, now runs on the Vercel AI Gateway. With more than 1 million developers and 4 million installations, Cline brings an AI coding partner directly into the development environment, grounded in the values of openness and transparency. To support that mission at scale, the team needed infrastructure that matched those principles: fast, reliable, and built on open standards. Read more

Vercel Knowledge Base

December 16, 2025

Vercel Knowledge Base is a new home for guides, tutorials, and best practices for developers building on Vercel, including how to: Build agentic apps with the AI cloud Host your Backend with your preferred framework Secure your apps with Vercel's built-in features You can use the Knowledge Base to find and explore guides for specific use cases with: Semantic AI search: describe what you're trying to achieve AI chat: ask our agent about a guide Filters: search guides by Vercel product or feature Read more

Export Observability query results to CSV or JSON

December 16, 2025

You can now export the results from your Observability queries as CSV or JSON files. This allows you to analyze, share, and process your Vercel observability data outside of the Vercel dashboard. Click the download icon on any query to export your query results instantly. This feature is available for all teams with Observability Plus. Try it out or learn more about Query. Read more

How to prompt v0

December 15, 2025

Working with v0 is like working with a highly skilled teammate who can build anything you need. v0 is more than just a tool, it’s your building partner. And like with any great collaborator, the quality of what you get depends on how clearly you communicate. Read more

Build smarter workflows with Notion and v0

December 15, 2025

Notion has become the trusted, connected workspace for teams. It's where your PRDs, specs, and project context live. v0 helps those teams turn ideas into dashboards, apps, and prototypes. Today, those workflows connect. You can now securely connect v0 to your Notion workspace, so everything it builds is grounded in your existing docs and databases. Wherever your team's knowledge lives in Notion, v0 can now build on top of it. Read more

Push notifications support on desktop and mobile

December 12, 2025

Push notifications are now available on both desktop and mobile, with support for all notification types. To start receiving push notifications from Vercel: Go to Notification Settings in the Vercel dashboard Enable the push notification channel for any notification type To allow mobile notifications on your phone: Open the Vercel Dashboard in your mobile browser Opt in to push notifications when prompted Try it out or learn more about notifications. Read more

Referer now available in runtime logs

December 12, 2025

For any request displayed on runtime logs in the Vercel dashboard, you can now view the referer (if any) for that request in the right hand details panel. This allows you to understand the source of that request and more easily debug issues. Try it out or learn more about Runtime Logs. Read more

Split Web Analytics data by any dimension

December 12, 2025

Web Analytics now allows you to split data across any dimension. You can now break down your Web Analytics data across any dimension, not just Flags and Flag Values. This update expands support to 11 dimensions, which are: Paths Routes Host names Countries Devices OS Referrers Flags Flag values Events names Event properties With dimension splits and filters, you can dig deeper into user activity and better understand how different segments are using your application. This feature is available to all Vercel users with the Web Analytics package installed. Web Analytics Plus subscribers also gain the enhanced capability of splitting data by UTM parameters. Try it out or learn more about Web Analytics. Read more

Add cache tags from Function responses, regardless of framework

December 12, 2025

You can now add one or more cache tags to your Function response by importing the addCacheTag function from @vercel/functions npm package. import { addCacheTag } from '@vercel/functions' Once the cached response has a tag associated with it, you can later invalidate the cache in one of several ways: Dashboard Settings Vercel CLI Function API REST API Available on all plans and all frameworks. Learn more about cache invalidation. Read more

React Server Components security update: DoS and Source Code Exposure

December 11, 2025

See the Security Bulletin for the latest updates. Summary Two additional vulnerabilities in React Server Components have been identified: a high-severity Denial of Service (CVE-2025-55184) and a medium-severity Source Code Exposure (CVE-2025-55183). These issues were discovered while security researchers examined the patches for the original React2Shell vulnerability. The initial fix was incomplete and did not fully prevent denial-of-service attacks for all payload types, resulting in CVE-2025-67779. Importantly, none of these new issues allow for Remote Code Execution. We created new rules to address these vulnerabilities and deployed them to the Vercel WAF to automatically protect all projects hosted on Vercel at no cost. However, do not rely on the WAF for full protection. Immediate upgrades to a patched version are required. Impact Denial of Service (CVE-2025-55184) A malicious HTTP request can be crafted and sent to any App Router endpoint that, when deserialized, can cause the server process to hang and consume CPU. Source Code Exposure (CVE-2025-55183) A malicious HTTP request can be crafted and sent to any App Router endpoint that can return the compiled source code of Serv

GPT 5.2 models now available on Vercel AI Gateway

December 11, 2025

You can now access OpenAI's latest GPT-5.2 models with Vercel's AI Gateway and no other provider accounts required. These models perform better than the GPT-5.1 model series, with noted improvements in professional knowledge work, coding, and long-context reasoning. Other highlights include fewer hallucinations, more accurate vision to interpret graphs and visualizations, strong complex front-end work capabilities, and better information retention working with long documents. There are 3 models available on AI Gateway: GPT-5.2 Chat (openai/gpt-5.2-chat) is the model used in ChatGPT, best suited for everyday work and learning. GPT-5.2 (openai/gpt-5.2) is for deeper work and complex tasks involving coding or long documents. GPT-5.2 Pro (openai/gpt-5.2-pro) is best suited for the most difficult questions and tasks with large amounts of reasoning. To use the GPT-5.2 models with the AI SDK, set the model to the respective model slug (noted above): AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, an

Node.js 24 LTS is now available on Sandbox

December 10, 2025

Vercel Sandbox now supports Node.js version 24. To run a Sandbox with Node.js 24, upgrade @vercel/sandbox to version 1.1.0 or above and set the runtime property to node24: Read our Sandbox documentation to learn more. Read more

Vercel launches partner certification

December 10, 2025

We're proud to introduce the inaugural cohort of Vercel Certified Solution Partners. These eleven industry-leading teams share our commitment to create a faster, more accessible, and more innovative web. This program is designed not only to validate partner expertise, but also to help customers confidently choose teams who understand their needs, technical requirements, and the experiences they aim to deliver. Through partner certification, customers are matched with teams proven to deliver exceptional outcomes with Next.js and Vercel, from ambitious redesigns and complex enterprise migrations to new product development. Read more

Inside Workflow DevKit: How framework integrations work

December 9, 2025

When we announced the Workflow Development Kit (WDK) at Ship AI just over a month ago, we wanted it to reflect our Open SDK Strategy, allowing developers to build with any framework and deploy to any platform. At launch, WDK supported Next.js and Nitro. Today it works with eight frameworks, including SvelteKit, Astro, Express, and Hono, with TanStack Start and React Router in active development. This post explains the pattern behind those integrations and how they work under the hood. Read more

FastAPI Lifespan Events are now supported on Vercel

December 9, 2025

Vercel now supports lifespan events for FastAPI apps. This allows you to define logic that can execute on startup and graceful shutdown—such as managing database connections or flushing external logs. Deploy FastAPI on Vercel or visit the FastAPI on Vercel documentation. Read more

Unified security actions dashboard

December 8, 2025

Vercel now provides a unified dashboard that surfaces any security issues requiring action from your team. When a critical vulnerability or security-related task is detected, the dashboard automatically groups your affected projects and guides you through the steps needed to secure them. This view appears as a banner whenever action is required, and can be accessed anytime through the dashboard search. Most CVEs are handled automatically through WAF rules and other protections, but when user action is needed, they will appear here. Automatic detection of security vulnerabilities that require user intervention - When the platform identifies a vulnerability or configuration that cannot be fully mitigated by Vercel’s autonomous protections, it’s surfaced here with clear instructions. Project grouping based on required actions - Current categories include unpatched dependencies, manual fix required, unprotected preview deployments. Additional groups will appear over time as new protections and checks are added. Support for both automated remediation - When possible, Vercel Agent offers one-click automated upgrades and PRs. Support for manual remediation - For cases requiring manual upd

Automated React2Shell vulnerability patching is now available

December 8, 2025

Vercel Agent now detects vulnerable packages in your project, and automatically generates pull requests with fixes to upgrade them to patched versions. Powered by Vercel's self-driving infrastructure, these auto-fix upgrades are available at no cost and help teams stay secure with minimal manual effort. Automatic detection of vulnerable React, Next.js, and related RSC packages Automatic PR creation Full execution and verification of updates inside isolated Sandbox environments Preview links generated with PR, to manually validate updates About React2Shell React2Shell (CVE-2025-55182) is a critical remote code execution vulnerability in React Server Components that affects React 19 and frameworks that use it like Next.js. Specially crafted requests can trigger unintended code execution if your application is running a vulnerable version. Immediate upgrades are required for all projects using affected React and Next.js releases. Get the latest updates on React2Shell or view the new dashboard here. Read more

Rust runtime now in public beta for Vercel Functions

December 8, 2025

Today, we are launching first-class support for the Rust runtime beta. This new release of native support, as an evolution of the community Rust runtime, brings the full benefits of Vercel Functions, including Fluid compute (with HTTP response streaming and Active CPU pricing) and an increased environment variable limit from 6KB to 64KB. Rust deployments automatically integrate with Vercel's existing logging, observability, and monitoring systems. To get started, create a Cargo.toml file and a handler function like in the example below: Deploy to Vercel today with one of our starter templates Rust Hello World and Rust Axum, or read more in the Function docs. Read more

Rewrites and redirects now available in runtime logs

December 5, 2025

Vercel users can now view requests that make rewrites or redirects directly in the Vercel dashboard in runtime logs. By default, these requests are filtered out on the Runtime Logs page. To view these requests on the Logs page, you can filter for Rewrites or Redirects in the Resource dropdown. Rewrites: shows the destination of the rewrite Redirects: shows the redirect status code and location This feature is available to all users. Try it out or learn more about runtime logs. Read more

New deployments of vulnerable Next.js applications are now blocked by default

December 5, 2025

Any new deployment containing a version of Next.js that is vulnerable to CVE-2025-66478 will now automatically fail to deploy on Vercel. We strongly recommend upgrading to a patched version regardless of your hosting provider. Learn more This automatic protection can be disabled by setting the DANGEROUSLY_DEPLOY_VULNERABLE_CVE_2025_66478=1 environment variable on your Vercel project. Learn more Read more

Introducing Platform Elements

December 5, 2025

As part of the new Vercel for Platforms product, you can now use a set of prebuilt UI blocks and actions to add functionality directly to your application. An all-new library of production-ready shadcn/ui components and actions help you launch (and upgrade) quickly. Blocks: custom-domain: Domain configuration with DNS validation and real-time verification deploy-popover: Deployment interface with status and history dns-table: DNS records in a copyable format claim-deployment: Ownership verification flow report-abuse: Abuse reporting form Actions: add-custom-domain: Add and verify domains programmatically deploy-files: Deploy files to any project You can install Platforms components with the Vercel Platforms CLI. For example: Start building with Platform Elements using our Quickstart for Multi-Tenant or Multi-Project platforms. Read more

Introducing Vercel for Platforms

December 5, 2025

You can now build platforms with the new Vercel for Platforms product announced today, making it easy to create and run customer projects on behalf of your users. Two platform modes are available: Multi-Tenant and Multi-Project, allowing you to deploy with a single codebase or many, across any number of domains. Multi-Tenant Platforms Run a single codebase that serves many customers with: Wildcard domains (*.yourapp.com) with automatic routing and SSL. Custom domain support via SDK, including DNS verification and certificate management. Routing Middleware for hostname parsing and customer resolution at the edge. Single deployment model: deploy once, changes apply to all tenants. Add custom domains to your app in seconds: Multi-Project Platforms Create a separate Vercel project per customer with: Programmatic project creation with the Vercel SDK. Isolation of builds, functions, environment variables, and settings per customer. Support for different frameworks per project. Deploy your customer's code into isolated projects in seconds: Today we are also introducing Platform Elements, a new library to make building on platforms easier. Start building with our Quickstart for Multi-Tenan

GPT 5.1 Codex Max now available on Vercel AI Gateway

December 5, 2025

You can now access OpenAI's latest Codex models, GPT-5.1 Codex Max with Vercel's AI Gateway and no other provider accounts required. Using a process called compaction, GPT-5.1 Codex Max has been trained to operate across multiple context windows and on real-world software engineering tasks. GPT-5.1 Codex Max is faster and more token efficient compared to previous Codex models, optimized for long-running coding tasks, and can maintain context and reasoning over long periods without needing to start new sessions. To use GPT-5.1 Codex Max with the AI SDK, set the model to openai/gpt-5.1-codex-max. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

React2Shell Security Bulletin

December 5, 2025

CVE-2025-55182 is a critical vulnerability in React that requires immediate action. Next.js and other frameworks that React are affected. Read the bulletin and act now. Read more

Domains must now be managed at the team level

December 4, 2025

Managing domains at the account level is no longer supported. Domains must now be managed at the team level, which simplifies access control, collaboration, and unified billing. Domains that are currently linked to accounts will continue to resolve, serve traffic, and renew as usual, but any changes will require moving the domain to a team. When viewing an account-level domain, you'll now be prompted to select a destination team to transfer your domain and all project domains, DNS records, and aliases will move to that team and continue to work after the move. Read more

New npm package for automatic recovery of broken streaming markdown

December 3, 2025

Remend is a new standalone package that brings intelligent incomplete Markdown handling to any application. Previously part of Streamdown's Markdown termination logic, Remend is now a standalone library (npm i remend) you can use in any application. Why it matters AI models stream Markdown token-by-token, which often produces incomplete syntax that breaks rendering. For example: Unclosed fences Half-finished bold/italic markers Unterminated links or lists Without correction, these patterns fail to render, leak raw Markdown, or disrupt layout: Remend automatically detects and completes unterminated Markdown blocks, ensuring clean, stable output during streaming. As the stream continues and the actual closing markers arrive, the content seamlessly updates, giving users a polished experience even mid-stream. It works with any Markdown renderer as a pre-processor. For example: Remend powers the markdown rendering in Streamdown and has been battle-tested in production AI applications. It includes intelligent rules to avoid false positives and handles complex edge cases like: Mathematical expressions with underscores in LaTeX blocks Product codes and variable names with asterisks/undersc

Summary of CVE-2025-55182

December 3, 2025

See the React2Shell security bulletin for the latest updates. Summary A critical-severity vulnerability in React Server Components (CVE-2025-55182) affects React 19 and frameworks that use it, including Next.js (CVE-2025-66478). Under certain conditions, specially crafted requests could lead to unintended remote code execution. We created new rules to address this vulnerability and quickly deployed to the Vercel WAF to automatically protect all projects hosted on Vercel at no cost. However, do not rely on the WAF for full protection. Immediate upgrades to a patched version are required. We also worked with the React team to deliver recommendations to the largest WAF and CDN providers. We still strongly recommend upgrading to a patched version regardless of your hosting provider. Impact Applications using affected versions of the React Server Components implementation may process untrusted input in a way that allows an attacker to perform remote code execution. The vulnerability is present in versions 19.0, 19.1.0, 19.1.1, and 19.2.0 of the following packages: : react-server-dom-parcel (19.0.0, 19.1.0, 19.1.1, and 19.2.0) react-server-dom-webpack (19.0.0, 19.1.0, 19.1.1, and 19.2.0)

Vercel Agent can now install Web Analytics and Speed insights for you

December 3, 2025

Vercel Agent can now automatically update your codebase and submit a PR to add Web Analytics and Speed Insights to your project. Vercel Agent analyzes your project configuration and connected GitHub repository, installs the relevant package, adds the relevant code snippet, and creates a pull request with the proposed changes. To have Vercel Agent install Web Analytics or Speed Insights to your project: Go to the Analytics or Speed Insights page of the dashboard. Enable the feature. Click Implement to start Vercel Agent. Review the pull request and merge when ready. Once the pull request is merged and deployed, tracking starts automatically. Vercel Agent installations are now available in Public Beta for all teams. Try it out for Web Analytics or Speed Insights. Read more

Nova 2 Lite now available on Vercel AI Gateway

December 3, 2025

You can now access Amazon's latest model Nova 2 Lite via Vercel's AI Gateway with no other provider accounts required. Nova 2 Lite is a reasoning model for everyday workloads that can process text, images, and videos to generate text. To use Nova 2 Lite, set model to amazon/nova-2-lite in the AI SDK. Extending thinking is disabled by default. To enable reasoning for this model, set maxReasoningEffort in the providerOptions. The reasoning content is redacted and displays as such, but users are still charged for these tokens. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Read the docs, view the AI Gateway model leaderboard, or use the model directly in our model playground. Read more

Billions of requests: Black Friday-Cyber Monday 2025

December 2, 2025

Every year, Black Friday and Cyber Monday reveal how people shop, browse, and discover products at global scale. For Vercel, the weekend doesn’t require a different operating mode. The platform behaves the same way it does every day, only with higher traffic volume. A live dashboard showed the traffic as it played out. This year, traffic reached more than 115.8 billion total requests, reflecting 33.6% year-over-year growth with consistent performance throughout the events. The traffic shape told a familiar story. Requests dipped on Thanksgiving as people stepped away from screens, then surged on Black Friday, stayed elevated through the weekend, and built into a second wave on Cyber Monday. These rhythms played out across every major geography, and the platform adapted continuously without configuration changes or manual intervention. Below is a snapshot of what the weekend looked like. Key metrics from the weekend 115,836,126,847 total requests - Global traffic delivered with consistent performance. 518,027 peak requests per second - Traffic delivered at peak demand 6,120,247 deployments - New versions of applications shipped 24,086,391 AI Gateway requests - AI routing kept respon

Investing in the Python ecosystem

December 2, 2025

The team behind Gel Data is joining Vercel to help us invest in the Python ecosystem. Led by Python core developer Yury Selivanov and contributor Elvis Pranskevichus, they will bring world-class support for Python on the AI Cloud. Read more

Deploy steps are now up to 21% faster

December 2, 2025

On average, the deploy step is now 17% faster, reducing total time to go live by 1.67 seconds. Projects with a large number of functions will see even greater improvements, with up to 2.8 seconds saved on average. During the deploy step, Vercel uploads static assets, provisions and uploads like Vercel Functions, processes routing metadata, and prepares the deployment to receive traffic. This phase is now faster due to reduced idle time and increased concurrency across these operations. Check out the documentation to learn more about builds. Read more

Mistral Large 3 now available on Vercel AI Gateway

December 2, 2025

You can now access Mistral's latest model Mistral Large 3 via Vercel's AI Gateway with no other provider accounts required. Mistral Large 3 is Mistral's most capable model to date. It has a sparse mixture-of-experts architecture with 41B active parameters (675B total), and is Mistral’s first mixture-of-experts model since the Mixtral series. To use Mistral Large 3, set model to mistral/mistral-large-3 in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Read the docs, view the AI Gateway model leaderboard, or use the model directly in our model playground. Read more

AWS Databases coming to the Vercel Marketplace

December 1, 2025

We are expanding our partnership with AWS to make it faster for developers to build and scale with AWS infrastructure. On December 15th, Aurora PostgreSQL, Amazon DynamoDB, and Aurora DSQL will be available as native integrations in the Vercel Marketplace. These integrations bring the power and scalability of AWS databases directly into your Vercel workflow, so you can focus on shipping products, agents, and websites instead of configuring infrastructure. Read more

Trinity Mini model now available in Vercel AI Gateway

December 1, 2025

You can now access Arcee AI's latest model Trinity Mini via Vercel's AI Gateway with no other provider accounts required. Trinity Mini is an open weight MoE reasoning model with 26B parameters (3B active) trained end-to-end in the U.S. To use Trinity Mini, set model to arcee-ai/trinity-mini in the AI SDK. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Read the docs, view the AI Gateway model leaderboard, or use the model directly in our model playground. Read more

DeepSeek V3.2 models now available in Vercel AI Gateway

December 1, 2025

You can now access DeepSeek's latest models, DeepSeek V3.2 and DeepSeek V3.2 Speciale, via Vercel's AI Gateway with no other provider accounts required. DeepSeek V3.2 supports combined thinking and tool use, handling agent-style operations (tool calls) in both reasoning and non-reasoning modes. DeepSeek V3.2 Speciale is optimized for maximal reasoning performance, and is suited for complex task use cases but requires higher token usage and does not support tool use. To use the DeepSeek V3.2 models, set model to the following in the AI SDK: Non-thinking: deepseek/deepseek-v3.2 Thinking: deepseek/deepseek-v3.2-thinking Speciale: deepseek/deepseek-v3.2-speciale AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Read the docs, view the AI Gateway model leaderboard, or use DeepSeek V3.2 models directly in our model playground. Read more

Runtime logs now appear in the dashboard 6x faster

December 1, 2025

We've improved our logs infrastructure resulting in significantly better performance when interacting with logs on Vercel: Logs appear up to 6× faster in the dashboard, with 90% of entries visible within 5 seconds of execution. These improvements make Live Mode a more responsive experience. Filtering and querying Runtime Logs is now up to 30% faster, with 80% of filter counts now resolving in under 1 second, to find what you need quicker. Learn more about Runtime Logs. Read more

Image-only models available in Vercel AI Gateway

November 28, 2025

You can now access image-only models via Vercel's AI Gateway with no other provider accounts required. In addition to multimodal models with image generation capabilities that are currently available in AI Gateway (e.g., GPT-5.1, Nano Banana Pro, etc.), these image-only models are exclusively for image generation. The models include: Black Forest Labs: FLUX.2 Flex: bfl/flux-2-flex FLUX.2 Pro: bfl/flux-2-pro FLUX.1 Kontext Max: bfl/flux-kontext-max FLUX.1 Kontext Pro: bfl/flux-kontext-pro FLUX 1.1 Pro Ultra: bfl/flux-pro-1.1-ultra FLUX 1.1 Pro: bfl/flux-pro-1.1 FLUX.1 Fill Pro: bfl/flux-pro-1.0-fill Google: Imagen 4.0 Generate 001: google/imagen-4.0-generate Imagen 4.0 Fast Generate 001: google/imagen-4.0-fast-generate Imagen 4.0 Ultra Generate 001: google/imagen-4.0-ultra-generate To use these models, set model to the corresponding slug from above in the AI SDK. These models support generateImage. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic

Sign in with Vercel now generally available

November 26, 2025

Sign in with Vercel is now generally available, enabling developers to add Vercel as a sign-in method to their apps and projects. You can create an app, configure its scopes, and start setting up sign-in directly from the Vercel dashboard, without having to manage users or sign-in methods yourself. Built on OAuth and OpenID, Sign in with Vercel allows you to: Sign users in to your apps and projects with their existing Vercel account. Fetch the user’s info, such as their name, email, and avatar. Receive ID tokens, access tokens, and refresh tokens for interacting with Vercel. Take a look at our example app to get started. Read more

Node.js 24 LTS is now generally available for builds and functions

November 25, 2025

Node.js version 24 is now available as a runtime for builds and functions using Node. To use version 24, go to Project Settings -> Build and Deployment -> Node.js Version and select 24.x. This is also the default version for new projects. This new version's highlights: V8 Engine Upgrade: Node.js 24 ships with the V8 JavaScript engine version 13.6, bringing performance enhancements and new JavaScript features such as Float16Array and Error.isError Global URLPattern API: Simpler URL routing and matching without the need for external libraries or complex regular expressions Undici v7: The built-in fetch API benefits from faster HTTP performance, improved HTTP/2 & HTTP/3 support, and more efficient connection handling npm v11: It comes with an updated version of npm, improving the compatibility with modern JavaScript packages The current version used is 24.11.0 and will be automatically updated, with only the major version version (24.x) being guaranteed. Read our Node.js runtime documentation to learn more. Read more

FLUX.2 Pro image model is now available on Vercel AI Gateway

November 25, 2025

You can now access the newest image model FLUX.2 Pro from Black Forest Labs via Vercel's AI Gateway with no other provider accounts required. FLUX.2 Pro is a newly trained base model designed for advanced visual intelligence, offering higher-resolution outputs (up to 4MP), improved knowledge of the real world, and precise control over lighting and spatial composition. It introduces multi-reference input, enhanced character and product consistency, exact color matching, and expanded control options compared to the FLUX.1 models. FLUX.2 Pro differs fundamentally from the other models with image generation capability currently available in AI Gateway. This model is a pure image-focused rectified-flow transformer model, in contrast with the multimodal LLMs already integrated. To use this model, set model to bfl/flux-2-pro in the AI SDK. This model supports generateImage. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Read the docs for more

.fast, .you, .talk, and .to TLDs now available on Vercel Domains

November 25, 2025

Vercel Domains now supports the TLDs .fast, .you, .talk, and .to.. Domains with these TLDs can now be purchased at vercel.com/domains, and you can also now transfer domains with these TLDs onto the Vercel platform for easy integration and use with projects and deployments. Try it here. Read more

Shai-Hulud 2.0 Supply Chain Compromise

November 24, 2025

Multiple npm packages from various web services were compromised through account takeover/developer compromise. A malicious actor was able to add a stealthy loader to the package.json file that locates the Bun runtime, silently installs, then executes a malicious script. Our investigation has shown that no Vercel environment was impacted and we are notifying a small set of customers with affected builds. Impact to Vercel Customers Vercel has taken immediate steps to address this for our customers. As an initial step, we reset the cache for projects that pulled in any of the vulnerable packages while we continue to investigate whether any loaders successfully ran. As of this publication, no Vercel-managed systems or internal build processes have been impacted. Preliminary analysis identified a limited set of Vercel customer builds referencing the compromised packages. Impacted customers are being contacted directly with detailed mitigation steps. We will continue to issue updates throughout our investigation. Read more

You can now configure advanced sampling rules for Vercel Drains

November 24, 2025

You can now configure advanced sampling rules when exporting data to a third-party observability tool when using Vercel Drains. Advanced sampling rules allow you to configure sampling rates for specific environments and path prefixes, providing more granular control over cost management. Vercel Drains is available to Pro and Enterprise teams. Advanced sampling rules can be configured on drains exporting logs or traces. Try it out or learn more about Vercel Drains. Read more

Claude Opus 4.5 now available in Vercel AI Gateway

November 24, 2025

You can now access Anthropic's latest model, Claude Opus 4.5, via Vercel's AI Gateway with no other provider accounts required. Claude Opus 4.5 is suited for demanding reasoning tasks and complex problem solving. This model has improvements in general intelligence and vision compared to previous iterations. It excels at difficult coding tasks and agentic workflows, especially those with computer use and tool use, and can effectively handle context usage and external memory files. Frontend coding and design are established strengths, particularly for developing real-world web applications. To use Claude Opus 4.5, set model to anthropic/claude-opus-4.5 in the AI SDK. There is a new effort parameter for this model. This parameter affects all types of tokens and controls the level of token usage when responding to a request. By default, effort is set to high and is independent of the thinking budget. To use it in AI Gateway with the AI SDK, set effort for the provider in providerOptions, as seen below in the example. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider up

Streamdown 1.6 is now available to run faster and ship less code

November 24, 2025

Vercel Streamdown 1.6 is now available with major improvements to performance, bundle size, and the authoring experience. Streamdown now runs faster and ships less code thanks to memoization, LRU caching, optimized string operations, and the removal of regexes. Several product enhancements include: Code Blocks, Mermaid, and Math components are now lazy-loaded with React.lazy() and Suspense, only loading when used. The code highlighting system has been rebuilt with a new tokenization approach that’s simpler, more efficient, and includes line numbers. A custom markdown renderer replaces React Markdown, giving Streamdown a lighter core and more room for future optimizations. Static Mode adds support for rendering markdown without streaming, ideal for blogs and other static use cases as it reduces streaming overhead. Mermaid blocks now support custom error components for handling parsing issues. Diagrams can be exported as SVG, PNG, or source code, and the fullscreen view includes zoom and pan controls (thanks to zhdzb). Update to Vercel Streamdown 1.6 today with npm i streamdown@latest or read more about Streamdown here. Read more

Open your Vercel dashboard from the Vercel CLI

November 24, 2025

You can now open your current project in the Vercel Dashboard directly from the command line using vercel open. This gives you quick access to your project without needing to navigate manually in your browser. Update vercel to 48.10.0 or newer with npm i -g vercel to give it a try. See docs for more. Read more

How we built the v0 iOS app

November 24, 2025

We recently released v0 for iOS, Vercel’s first mobile app. As a company focused on the web, building a native app was new territory for us. Our goal was to build an app worthy of an Apple Design Award, and we were open-minded on the best tech stack to get there. To that end, we built dozens of iterations of the product prior to our public beta. We experimented with drastically different tech stacks and UI patterns. We took inspiration from apps which speak the iPhone’s language, such as Apple Notes and iMessage. v0 had to earn a spot on your Home Screen among the greats. After weeks of experimentation, we landed on React Native with Expo to achieve this. We are pleased with the results, and our customers are too. In fact, the influx of messages from developers asking how the app feels so native compelled us to write a technical breakdown of how we did it. Table of contents How we built the v0 chat experience Building a composable chat Sending your first message Fading in the first assistant message Sending messages in an existing chat How we solved messages scrolling to the top Taming the keyboard Scrolling to end initially Floating composer Make it float Make it native Pasting im

Security through design: Creating the improved Firewall experience

November 24, 2025

At Vercel, we believe security should be intuitive, not intimidating. The best security tool is the one that's actually used. It should be clear, useful, and never in the way. But that's not always the norm. Security tooling can often feel like a tradeoff against shipping velocity. When UX is an afterthought, teams leave tools off or in "logging mode" forever, even when risks are high. That's why we've redesigned the Vercel Firewall experience from the ground up. The new UI helps you see more, do more, and feel confident in your app's resilience to attacks. Designing for every Vercel user The redesign started with listening. Users told us: I want to easily see active DDoS events I need more information on what the Firewall blocked I need a faster way to investigate traffic alerts or spikes Developers, SREs, and security teams all use the Firewall for maintenance and troubleshooting. They configure rules, monitor traffic, and respond to unusual activity. The new Firewall UI is designed for everyone using Vercel. It surfaces clear, actionable information, simplifies navigation, and helps teams resolve issues quickly when it matters most. A better way to see and secure your traffic Th

Workflow Builder: Build your own workflow automation platform

November 24, 2025

Today we're open-sourcing Workflow Builder, a complete visual automation platform powered by the Workflow Development Kit (WDK). The project includes a visual editor, execution engine, and infrastructure, giving you what you need to build your own workflow automation tools and agents. Deploy it to Vercel and customize it for your use case. Read more

Deployments can now require cryptographically-verified commits

November 24, 2025

Vercel now supports commit verification, letting you protect your deployments by requiring commits to be cryptographically verified before they’re deployed from GitHub. Enable it for GitHub-connected projects in your project settings. Learn more about commit signing and verification on GitHub or read more about the setting in our docs. Read more

Convex joins the Vercel Marketplace

November 24, 2025

Convex is now available on the Vercel Marketplace, giving developers an easy way to add a real-time backend to any Vercel project. You can create and connect a Convex project directly from the Vercel dashboard and get a fully configured backend without manual setup. With the new integration, you can: Provision a Convex project directly from Vercel dashboard Manage accounts and billing in one place Get real-time data sync with built-in caching and consistency Use Convex’s data model and functions alongside Vercel’s full developer workflow Install Convex from the Marketplace and start building with a fully connected backend in just a few clicks. Read more

Vercel Agent investigations now included in Observability Plus

November 21, 2025

Vercel Agent investigations are now included in Observability Plus, adding 10 investigations to every billing cycle at no extra cost to your subscription. Investigations help teams diagnose and resolve incidents faster, and run automatically on error alerts. When an alert flags suspicious activity, such as unexpected spikes in usage or errors, Vercel Agent investigates the issue, identifies the likely root cause, analyzes the impact, and suggests next steps for remediation. Teams can purchase Vercel Agent credits to run additional investigations. Investigations are public beta for Pro and Enterprise teams with Observability Plus. Try it out or learn more about Vercel Agent investigations. Read more

Vercel Open Source Program: Fall 2025 cohort

November 21, 2025

In April, we launched the Vercel Open Source Program to give maintainers the resources, credits, and support they need to ship faster and scale confidently. The first group joined through our spring 2025 cohort. Today we are welcoming the fall 2025 cohort. From AI-native apps and developer infrastructure to design systems and creative tooling, open-source builders continue to amaze us. Meet the creators and explore their projects. Read more

Self-driving infrastructure

November 21, 2025

AI has transformed how we write code. The next transformation is how we run it. At Vercel, we’re building self-driving infrastructure that autonomously manages production operations, improves application code using real-world insights, and learns from the unpredictable nature of production itself. Our vision is a world where developers express intent, not infrastructure. Where ops teams set principles, not individual configurations and alerts. Where the cloud doesn’t just host your app, it understands, optimizes, and evolves it. Read more

Grok 4.1 Fast models now available on Vercel AI Gateway

November 20, 2025

You can now access xAI's latest models, Grok 4.1 Fast Reasoning and Grok 4.1 Fast Non-Reasoning, via Vercel's AI Gateway with no other provider accounts required. These models have a 2M context window and are designed for agentic tool calling. Grok 4.1 Fast Reasoning is best suited for structured reasoning and agentic operations that require high accuracy, whereas Grok 4.1 Fast Non-Reasoning is tailored to speed. To use the Grok 4.1 Fast models in AI Gateway with the AI SDK, set model to xai/grok-4.1-fast-reasoning or xai/grok-4.1-fast-non-reasoning. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

You can now invalidate the CDN cache by providing a source image

November 20, 2025

Vercel Image Optimization dynamically transforms source images to reduce file size while maintaining high quality on the visitor's browser. You can now invalidate the CDN cache by providing a source image. This feature marks all transformed images derived from that source image as stale. The next request serves stale content instantly while revalidation happens in the background, with no latency impact for users. There are several ways to invalidate a source image: Dashboard Settings Vercel CLI Function API REST API In addition to invalidating by source image, you can also delete by source image if the origin is gone. Deleting the cache can increase latency while new content is generated, or cause downtime if your origin is unresponsive. We recommend you use with caution. This is available on all plans using the new image optimization price. Learn more about cache invalidation. Read more

Improved analytics experience now available on the Vercel Firewall

November 20, 2025

We have launched improvements to the Vercel Firewall UI, simplifying your application security monitoring and analysis. Vercel Firewall includes the System firewall and DDoS mitigations, Web Application Firewall, and Bot Management capabilities. The updated experience surfaces more information on security events and mitigations, and allows for easier event investigations, bringing together all security events analytics in one place. The updates include: An updated Overview page for a consolidated view of DDoS attacks, and activity across system rules, custom rules, and IP blocks. A new Traffic page that allows you to drill down into top sources of traffic (IPs, request paths, JA4 digests, ASN, user agents) and filter by actions (allowed, logged, denied, challenged, rate limited) Simplified UX for writing custom rules or queries so that you can take actions or do analysis without friction Learn more about the Firewall or visit the Firewall tab on your project to see the updates. Read more

Nano Banana Pro (Gemini 3 Pro Image) now available in the AI Gateway

November 20, 2025

You can now access Google's cutting edge image model, Nano Banana Pro (Gemini 3 Pro Image), via Vercel's AI Gateway with no other provider accounts required. Nano Banana Pro (Gemini 3 Pro Image) is designed to work for more advanced use cases than Nano Banana. This model introduces improvements specifically for professional and creative workflows, like the generation of diagrams with accurate labeling and integration of web search information for images with up-to-date information. Nano Banana Pro also supports higher resolution generation and higher multi-image input limits for better compositing. To use Nano Banana Pro in AI Gateway with the AI SDK, set model to google/gemini-3-pro-image. Note that this is a multi-modal model and therefore uses generateText for the actual image generation. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Read the AI Gateway docs for examples on how to use Nano Banana Pro to generate images, view the AI

Vercel collaborates with Google for Gemini 3 Pro Preview launch

November 18, 2025

The Gemini 3 Pro Preview model, released today, is now available through the Vercel AI Gateway and on v0.app. Thanks to Google, Vercel has been testing Gemini 3 Pro Preview across v0, Next.js, AI SDK, and Vercel Sandbox over the past several weeks. We've noticed the model has an increased focus on coding, multimodal reasoning, and tool use, though it's seen improvements across the board. From our testing, Gemini 3 Pro Preview delivers substantial improvements in instruction following and response consistency. It shows almost a 17% increase in correctness over its predecessor on our Next.js evals, putting in the top 2 models on the leaderboard. Read more

Gemini 3 Pro now available in Vercel AI Gateway

November 18, 2025

You can now access Google's latest model, Gemini 3 Pro, via Vercel's AI Gateway with no other provider accounts required. Gemini 3 Pro excels at challenging tasks involving reasoning or agentic workflows. In particular, the model improves on Gemini 2.5 Pro's performance in multi-step function calling, planning, reasoning over complex images/long documents, and instruction following. To use Gemini 3 Pro in AI Gateway with the AI SDK, set model to google/gemini-3.0-pro-preview. Gemini 3 Pro is a reasoning model, and you can specify the level of thinking. Include the providerOptions configuration with includeThoughts like the example below to enable reasoning text. AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Vercel now supports Build Commands for FastAPI and Flask

November 17, 2025

You can now can easily deploy FastAPI and Flask with custom Build Commands, expanding support for Python projects on Vercel. In addition to defining a Build Command in the project Settings dashboard, you can also define a build script in[tool.vercel.scripts]inside your pyproject.toml. This script will run after dependencies are installed, but before your application is deployed. Learn more about the Build Command for Python projects. Read more

Elysia can now be automatically deployed on Vercel

November 17, 2025

Elysia, a popular ergonomic TypeScript framework with end-to-end type safety, can now be deployed instantly on Vercel. When deployed, Vercel will now automatically identify your app is running Elysia and provision the optimal resources to run it efficiently. By default, Elysia will use Node. You can opt-in to the Bun runtime by adding the bunVersion line below to your vercel.json. Backends on Vercel use Fluid compute with Active CPU pricing by default, so you only pay for time where your code is actively using CPU. Deploy Elysia on Vercel, or visit the documentation for Elysia or Bun Runtime at Vercel. Read more

Bulk redirects are now generally available

November 13, 2025

Vercel now supports bulk redirects, allowing up to one million static URL redirects per project. This feature adds import options for formats like CSV and JSON, so teams can more easily manage large-scale migrations, fix broken links, handle expired pages, and more. To use bulk redirects, set the bulkRedirectsPath field in your vercel.json to a file or folder containing your redirects. These will be automatically imported at build time. This feature is available for Pro and Enterprise customers, and includes rates for additional capacity: Pro: 1,000 bulk redirects included per project Enterprise: 10,000 bulk redirects included per project Additional capacity: starts at $50/month per 25,000 redirects Get started with bulk redirects. Read more

GPT 5.1 Codex models now available in Vercel AI Gateway

November 13, 2025

You can now access OpenAI's latest Codex models, GPT-5.1 Codex and GPT-5.1 Codex mini with Vercel's AI Gateway and no other provider accounts required. These Codex models are optimized for long-running, agentic coding tasks and are able to maintain context and reasoning over longer sessions without degradation. To use these models with the AI SDK, set the model to openai/gpt-5.1-codex or openai/gpt-5.1-codex-mini: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

GPT 5.1 models now available in Vercel AI Gateway

November 13, 2025

You can now access OpenAI's latest models, GPT-5.1 Instant and GPT-5.1 Thinking, using Vercel's AI Gateway with no other provider accounts required. GPT-5.1 Instant offers improved instruction following, adaptive reasoning, and warmer, more conversational responses. GPT-5.1 Thinking builds on GPT-5 Thinking with dynamic performance tuning that prioritizes speed for simple tasks and deeper reasoning for complex ones. To use these models with the AI SDK, set the model to openai/gpt-5.1-instant or openai/gpt-5.1-thinking: AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover, and performance optimizations for higher-than-provider uptime. It includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Rollbar joins the Vercel Marketplace

November 12, 2025

Rollbar is now available as a native integration on the Vercel Marketplace, bringing real-time error monitoring and code-first observability directly into your Vercel workflow. With Rollbar, developers can automatically detect, track, debug, and resolve faster across deployments, connecting every issue back to the exact release and commit that introduced it. This helps teams move quickly while staying confident in production. In just a few clicks, you can: Manage accounts and billing in one place Connect Rollbar to one or many Vercel projects in minutes Automatically track deployments and tie errors to the specific revision that caused them Keep environments and source maps aligned across Rollbar and Vercel for clean, readable stack traces Install Rollbar from the Vercel Marketplace. Read more

Model fallbacks now available in Vercel AI Gateway

November 10, 2025

Vercel's AI Gateway now supports fallback models for when models fail or are unavailable. In addition to safeguarding against provider-level failures, model fallbacks can help with errors and capability mismatches between models (e.g., multimodal, tool-calling, etc.). Fallback models will be tried in the specified order until a request succeeds or no options remain. Any error, such as context limits, unsupported inputs, or provider outages, can trigger a fallback. Requests are billed based on the model that completes successfully. This example shows an instance where the primary model does not support multimodal capabilities, falling back to models that do. To use, specify the model fallbacks in models within providerOptions: To have pre-defined provider routing in addition to model routing, specify both models and providers (order or only) in providerOptions: AI Gateway also includes built-in observability, Bring Your Own Key support, and supports OpenAI-compatible API. Read more

Support for TanStack Start

November 10, 2025

Vercel detects and supports TanStack Start applications, a full-stack framework powered by TanStack Router for React and Solid. Create a new TanStack Start app or add nitro() to vite.config.ts in your existing application to easily deploy your projects: TanStack Start apps on Vercel use Fluid compute with Active CPU pricing by default. This means your TanStack Start app will automatically scale up and down based on traffic, and you only pay for what you use, not for idle function time. Visit the TanStack Start on Vercel documentation to learn more Read more

Vercel: The anti-vendor-lock-in cloud

November 10, 2025

Vendor lock-in matters when choosing a cloud platform. Cloud platforms can lock you in by requiring you to build against their specific primitives. Vercel takes a different approach: you write code for your framework, not for Vercel. On AWS, you configure Lambda functions, NAT Gateways, and DynamoDB tables. On Cloudflare, you write Workers, use KV stores, Durable Objects, and bind services with Worker Service Bindings. These primitives only exist with that vendor, which means migrating to another platform requires rewriting your application architecture. Too often, cloud platforms make these choices for you. They define proprietary primitives, APIs, and services that pull your code deeper into their ecosystem until leaving becomes impractical. At Vercel, we believe the opposite approach creates better software and a healthier web. We want developers to stay because they want to, not because they have to. That means building open tools, embracing standards, and ensuring your code remains portable no matter where it runs. So Vercel works differently. It interprets your framework code and provisions infrastructure automatically. Your application does not need to know it runs on Vercel

How Nous Research used BotID to block automated abuse at scale

November 7, 2025

AI lab Nous Research made Hermes, their open-source language model, free for one week to increase accessibility. Within days, automated scripts overwhelmed the service with fake accounts performing high-volume inference requests across thousands of accounts to bypass rate limits. Despite having Cloudflare Turnstile in place, bulk signups continued. The abuse led to wasted inference compute and inflated identity provider bills. After the promotion ended, Nous realized that before reintroducing any kind of free tier, it needed a stronger layer of bot protection. Read more

Vercel now supports post-quantum cryptography

November 7, 2025

HTTPS connections to the Vercel network are now secured with post-quantum cryptography. Most web encryption today could be broken by future quantum computers. While this threat isn’t immediate, attackers can capture encrypted traffic today and decrypt it later as quantum technology advances. Vercel now supports post-quantum encryption during TLS handshakes, protecting applications against these future risks. Modern browsers will automatically use it with no configuration or additional cost required. Read more about encryption and how we secure your deployments. Read more

AI domain search now available via Vercel Domains

November 7, 2025

You can now search for domains on Vercel using AI-powered smart search. Press space in the search bar to enter smart search mode. This mode uses AI to suggest domain names based on your input. In smart search, you can: Click a domain name to generate similar suggestions. Search across all supported TLDs for that name. Try it at vercel.com/domains. Read more

Vercel Sandbox CLI is now available

November 7, 2025

We’ve introduced the Vercel Sandbox CLI, a command-line interface for managing isolated compute environments. Built on the familiar Docker CLI model, developers can now: Create and run sandboxes for Node.js (node22) or Python (python3.13) workloads. Execute commands inside existing sandboxes. Copy files between local and remote environments. List, stop, and remove sandboxes across projects and teams. Run interactively with support for --tty, --interactive, and --publish-port for port forwarding. Automate workflows via authentication tokens, environment variables, and timeouts. Full reference now available in the Sandbox CLI docs. Read more

What we learned building agents at Vercel

November 6, 2025

Agents present incredible promise for increased productivity and higher quality outcomes in enterprises. Companies are already using them to streamline customer support, code reviews, and sales operations. When building custom internal agents, the challenge isn't whether AI can create value, it's identifying the problems it's ready to solve today, at a cost that makes sense for the business. At Vercel, we are going through the same AI transformation as our customers. We use our own products to build agents that help us move faster and spend more time on meaningful work. After months of experimentation, we’ve turned our learnings into a repeatable methodology for finding and investing in AI projects that have the highest likelihood of creating significant business impact. Read more

Moonshot AI's Kimi K2 Thinking models are now available on Vercel AI Gateway

November 6, 2025

You can now access Moonshot AI's latest and most powerful thinking models, Kimi K2 Thinking and Kimi K2 Thinking Turbo, using Vercel's AI Gateway with no other provider accounts required. Kimi K2 Thinking is oss and excels at deep reasoning, handling up to 200–300 sequential tool calls, and achieves top results on benchmarks for reasoning and coding. Kimi K2 Thinking Turbo is a high speed version of Kimi K2 Thinking and is best suited for scenarios requiring both deep reasoning and low latency. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher-than-provider-average uptime. To use it with the AI SDK, set the model to moonshotai/kimi-k2-thinking or moonshotai/kimi-k2-thinking-turbo: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

CVE-2025-52662: XSS on Nuxt DevTools

November 6, 2025

A medium-severity security vulnerability in Nuxt DevTools was responsibly disclosed, and has been fixed for version 2.6.4. This issue may have allowed Nuxt auth token extraction via XSS under certain configurations. Nuxt DevTools users are encouraged to upgrade to the latest version. Read more details below. Summary A vulnerability chain in Nuxt DevTools allows remote code execution in development environments through a combination of cross-site scripting (XSS), authentication token exfiltration, and path traversal. Impact The vulnerability exists in the DevTools authentication page where error messages are rendered without proper sanitization, enabling DOM-based XSS. An attacker can exploit this to steal authentication tokens and leverage a path traversal vulnerability in the WebSocket message handler to write arbitrary files outside the intended directory, leading to remote code execution when configuration files are overwritten. Resolution The XSS was resolved by displaying errors as textContent instead of innterHTML in: Nuxt DevTools 2.6.4 Workarounds Avoid publicly exposing Nuxt DevTools or running Nuxt in production using Dev mode Credit Thanks to @yuske for responsible discl

CVE-2025-48985: Input Validation Bypass on AI SDK

November 6, 2025

A low-severity security vulnerability in Vercel's AI SDK was responsibly disclosed, and has been fixed for 5.0.52, 6.0.0-beta.* The issue may have allowed users to bypass filetype whitelists when uploading files. Vercel customers are encouraged to upgrade to the latest version. Read more details below. Summary Vulnerability in Vercel's AI SDK prompt conversion pipeline where improper URL-to-data mapping allows attackers to substitute arbitrary downloaded bytes for different supported URLs within the same prompt. The vulnerability occurs in the convert-to-language-model-prompt.ts file when filtering downloaded results could cause index misalignment between the downloadedFiles array and the original plannedDownloads array. Impact When processing mixed supported and unsupported URLs, the filtering operation removes null entries for supported URLs, causing the remaining downloaded data to be incorrectly associated with different URL keys. This results in bytes from an unsupported URL being mapped to a supported URL slot, allowing attackers to inject arbitrary content while bypassing URL-based trust and content validation mechanisms. This affects most methods that accepted images or fil

Skew Protection max age now supports the full deployment lifetime

November 6, 2025

Skew Protection helps ensure that requests for a user's session are consistently routed to the same deployment, even when new versions are being rolled out. You can now configure your project's Skew Protection max age to persist for the entire lifetime of your deployments. This removes the previous limits of 12 hours on Pro and 7 days on Enterprise. Set the value to any duration less than or equal to your project's Deployment Retention policy. Learn more about Skew Protection and enable it in your project. Read more

Edge Config reads and writes now billed per unit

November 6, 2025

Update (July 2026): Edge Config is now Global Config. The same API is available as @vercel/global-config; existing @vercel/edge-config projects keep working unchanged. Edge Config Reads and Writes are moving from package-based to per-unit pricing on the Pro plan. You’ll continue paying the same effective rates, but at the start of your next billing cycle you’ll now be billed per unit to align costs directly with your usage. The new rates are: Edge Config Reads: $0.000003 per read (prev. $3 per 1M reads) Edge Config Writes: $0.01 per write (prev. $5 per 500 writes) Per‑unit billing scales more smoothly across team sizes and usage patterns. It also helps teams on Pro use Edge Config without immediately consuming a large portion of the included monthly usage credit. Get started or learn more about Edge Config. Read more

How AI Gateway runs on Fluid compute

November 6, 2025

AI Gateway is a Node.js service for connecting to hundreds of AI models through a single interface. It processes billions of tokens per day. The secret behind that scale is Fluid. Read more

Free Vercel BotID Deep Analysis through January 15

November 5, 2025

BotID Deep Analysis, Vercel’s advanced bot protection system, will be free for all Pro and Enterprise customers from November 5 to January 15, 2026. BotID is an invisible CAPTCHA to stop advanced, human-like bots from attacking high-value endpoints like registrations, AI invocations, and checkouts. Deep Analysis, our most advanced solution, uses thousands of telemetry points for real-time client-side checks. To participate, visit the Bot Management section in the Firewall dashboard and opt in. BotID usage will not be billed during this period. Regular billing resumes on January 16. Read more

Build and deploy data applications on Snowflake with v0

November 4, 2025

We're announcing an integration with Snowflake for v0. With this, you can connect v0 to Snowflake, ask questions about your data, and build data-driven Next.js applications that deploy directly to Snowflake. The application and authentication are managed through Vercel's secure vibe coding architecture, while compute runs on Snowflake's secure and governed platform, ensuring that your data never leaves your Snowflake environment. Sign up for the waitlist to get notified when it's ready for testing. Read more

Route build traffic through Static IPs

November 4, 2025

You can now choose whether build traffic, such as calls to external APIs or CMS data sources during the build process, routes through your Static IPs. To enable this, go to your Project Settings → Connectivity → toggle "Use static IPs for builds." By default, this setting is disabled. When enabled, both build and function traffic will route through Static IPs and count toward Private Data Transfer usage. This is available to all teams using Static IPs. Try it out or learn more here. Read more

Redirects and rewrites now available in Observability

November 3, 2025

Improved observability into redirects and external rewrites is now available to all Vercel customers. External rewrites forward requests to APIs or websites outside your Vercel project, effectively allowing Vercel to function as a reverse proxy or standalone CDN. Customers on all plans get new views that offer visibility into key rewrite metrics: Total external rewrites External rewrites by hostnames Customers on Pro and Enterprise plans can upgrade to Observability Plus to get: Connection latency to external host Rewrites by source/destination paths Routes and paths for redirect location Drains have also been updated to support the following: Redirects External Rewrites Cached External Rewrites View external rewrites or learn more about Observability. Read more

Caching details now available in Runtime Logs

October 31, 2025

You can now view more details on how Vercel's CDN globally serves cached content to users as quickly as possible. In the right-hand panel of the Runtime Logs page, we now list: Cache key: A unique identifier for a specific version of a cached page Cache tags: Tags associated with the cached data Revalidation reason: If a revalidation took place, the reason why the content was being revalidated (time-based, tag-based, or deployment-based) This is available to all Vercel users at no additional cost. Try it out or learn more about Runtime Logs. Read more

BotID Deep Analysis catches a sophisticated bot network in real-time

October 31, 2025

On October 29 at 9:44am, BotID Deep Analysis detected an unusual spike in traffic patterns across one of our customer's projects. Traffic increased by 500% above normal baseline. What made this particularly interesting wasn't just the volume increase. The spike appeared to be coming from legitimate human users. Our team immediately began investigating and reached out to the customer to discuss what appeared to be an influx of bot traffic cleverly disguised as human activity. But before we could even complete that conversation, something remarkable happened: Deep Analysis, powered by Kasada’s machine learning backend, had already identified the threat and adapted to correctly classify it. Read more

Vercel Agent can now run AI investigations

October 31, 2025

Vercel is reimagining incident response for the agentic age. At Ship AI, we launched Vercel Agent Investigations in Public Beta, a new skill of Vercel Agent that automatically detects issues in your application, conducts root cause analysis, and provides actionable remediation plans to resolve incidents faster. Vercel Agent already helps teams with AI-powered code reviews. Now, it's expanding to help with incident response. By combining our newly-released anomaly alerts with investigations, we're improving how development teams respond to and resolve production issues. Read more

Zero-configuration support for Fastify

October 31, 2025

Vercel now supports Fastify applications, a web framework highly focused on providing the best developer experience with the least overhead and a powerful plugin architecture, with zero-configuration. Backends on Vercel use Fluid compute with Active CPU pricing by default. This means your Fastify app will automatically scale up and down based on traffic, and you only pay for what you use. Deploy Fastify on Vercel or visit the Fastify on Vercel documentation Read more

Microfrontends now generally available

October 31, 2025

Microfrontends support on Vercel is now generally available, enabling you to split large applications into smaller, independently deployable units that render as one cohesive experience for users. Each team can use their own framework and release cadence, while Vercel handles edge composition and routing for a seamless user experience. Since the public beta, we've improved domains routing support, added microfrontends to Observability, and simplified onboarding. Vercel is serving nearly 1 billion microfrontends routing requests per day, and over 250 teams, including Cursor, The Weather Company, and A+E Global Media are already deploying microfrontends. Pricing Included: 2 microfrontend projects Additional projects: $250 per project per month (available on Pro and Enterprise plans) Routing: $2 per million routing requests Pricing starts today for new projects and on November 30, 2025 for existing ones. If already used, the 3rd project for existing microfrontends users during the beta will continue to be free. Get started with microfrontends, clone one of our examples, or learn more in our documentation. Read more

Vercel achieves TISAX AL2 compliance to serve automotive partners

October 29, 2025

We’re proud to share that Vercel has successfully completed its assessment for the Trusted Information Security Assessment Exchange (TISAX) Level 2 (AL2). This milestone reinforces our commitment to delivering secure, reliable, and compliant infrastructure to our global customers, particularly those in the automotive and manufacturing sectors that require specific security and data protections. This achievement builds on our broader compliance program, which includes ISO/IEC 270001:2022, SOC 2 Type II, PCI DSS, HIPAA and more. Read more

OpenAI's GPT-OSS-Safeguard-20B now available in Vercel AI Gateway

October 29, 2025

You can now access OpenAI's latest open source model, GPT-OSS-Safeguard-20B using Vercel's AI Gateway with no other provider accounts required. GPT-OSS-Safeguard-20B is a fine-tuned version of its general-purpose GPT-OSS model, designed for developers to implement custom, policy-driven content moderation. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK, set the model to openai/gpt-oss-safeguard-20b: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Vercel achieves TISAX AL2

October 29, 2025

Vercel has achieved Trusted Information Security Assessment Exchange (TISAX) Assessment Level 2 (AL2), a security standard widely adopted across the automotive and manufacturing industries to evaluate information security and the use of cloud services within the supply chain. Customers can access Vercel’s TISAX assessment results directly through the ENX portal. To view the assessment details: Sign in to your account on the ENX portal Search for Vercel or look up the following details: Assessment ID: AMR06H-1 Scope ID: SYN3TM Read our blog post to learn more about TISAX and automotive compliance on Vercel. Read more

Bun runtime on Vercel Functions

October 28, 2025

We now support Bun as a runtime option for Vercel Functions, available in Public Beta. You can choose between Node.js and Bun for your project, configuring runtime behavior based on workload. We're working closely with the Bun team to bring this capability to production. This flexibility allows you to choose what works best for your use case. Use Node.js for maximum compatibility or switch to Bun for compute-intensive applications that benefit from faster execution. Through internal testing, we've found that Bun reduced average latency by 28% in CPU-bound Next.js rendering workloads compared to Node.js. These gains come from Bun's runtime architecture, built in Zig with optimized I/O and scheduling that reduce overhead in JavaScript execution and data handling. Read more

MiniMax M2 now available for free in Vercel AI Gateway

October 28, 2025

You can now access MiniMax's latest open source model, MiniMax M2 using Vercel's AI Gateway with no other provider accounts required. The model is free to use until Nov 7th, 2025. Focused on agentic use, Minimax M2 is very efficient to serve, with only 10B active parameters per forward pass. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK, set the model to minimax/minimax-m2: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Bun runtime now in Public Beta for Vercel Functions

October 28, 2025

The Bun runtime is now available in Public Beta for Vercel Functions. You can choose between Node.js and Bun as your project runtime, selecting the best option for your workload. Benchmarks show Bun reduced average latency by 28% for CPU-bound Next.js rendering compared to Node.js. To use Bun in Vercel Functions, set the runtime globally in your project's vercel.json: We currently support the following frameworks, with more on the way: Next.js Hono Express Nitro Bun supports TypeScript with zero configuration. Here's an example with Hono: Or get started with one of our starter templates: Express on Bun Hono on Bun Bun deployments automatically integrate with Vercel's existing logging, observability, and monitoring systems. See benchmarks in our blog post, or read the docs to learn more. Read more

David Totten Joins Vercel to Lead Global Field Engineering

October 27, 2025

The next era of enterprise technology will be defined by shortening the distance between idea and impact. At Vercel we help companies move faster, turning velocity into measurable business outcomes by anticipating customer needs and owning results end-to-end. To help us lead in this new era, David Totten is joining Vercel as VP of Global Field Engineering. David brings two decades of experience building and scaling technical teams, and he'll lead a unified organization that includes Sales Engineering, Developer Success, Professional Services, and Customer Support Engineering. Read more

Vercel Ship AI 2025 recap

October 27, 2025

Earlier this year we introduced the foundations of the AI Cloud: a platform for building intelligent systems that think, plan, and act. Last week at Ship AI, we showed what comes next. We launched new SDKs, infrastructure, and open source templates that make building production-ready agents as intuitive as building a standard feature. You can now define, deploy, and operate intelligent workflows on the same platform that powers your apps. Whether you're building your first agent or delivering it to millions of users, these releases make AI development as accessible and scalable as web development. Read more

AI Chat now available on Vercel docs

October 24, 2025

We're excited to announce that AI Chat is now live within the Vercel docs along with a subtle design overhaul. You can now get instant, conversational assistance directly on all docs pages. Ask about anything on the Vercel docs Load specific pages as context to get page-aware answers Copy chat as Markdown for sharing or saving conversations How to use it Go to vercel.com/docs Click the Ask AI button at the top right corner in the header of any Vercel docs page to start asking questions Use the Ask AI about this page at the top of each docs page to load a page as context for focused learning, or copy the conversation as Markdown to share with your team. Read more

Manage Next.js Server Actions in the Vercel Firewall

October 24, 2025

The Vercel Firewall and Observability Plus has first-class support for Server Actions. Starting with Next.js 15.5, customers can now configure custom rules targeting specific server action names. In the example below, you can rate limit app/auth/actions.ts#getUser actions to 100 requests per minute per IP address. Server Action Name is available in the Firewall for all plans at no additional cost. Read the docs to learn more. Read more

Zero-config backends on Vercel AI Cloud

October 23, 2025

The same ease of use you expect from Vercel, now extended to your backends. Since we introduced the AI Cloud at Vercel Ship, teams have been building AI applications that go beyond simple prompt-to-response patterns. These apps orchestrate multi-step workflows, spawn sub-agents, and run processes that take hours or days. They need backends that process data, run inference, and respond to real-time events. You can now deploy the most popular Python and TypeScript backend frameworks with zero configuration. Vercel reads your framework and automatically provisions the infrastructure to run it. Read more

Introducing Vercel Agent: Your new Vercel teammate

October 23, 2025

We're launching Vercel Agent, an AI teammate for your development workflow. Vercel Agent uses AI, deep platform expertise, your application code, and telemetry data from across Vercel to help you ship faster with higher quality. Starting today, Vercel Agent is available in Public Beta with two core skills: Code Review and Investigations. Read more

Introducing AI agents & services on the Vercel Marketplace

October 23, 2025

The Vercel Marketplace now includes a dedicated AI Agents & Services category, making it easier for developers to integrate AI-powered automation, observability, and infrastructure directly into their projects. This category introduces native support for agentic integrations with unified authentication, provisioning, and billing across providers, all within the Vercel platform. Agents - Off-the-shelf agents that reason and act on your behalf CodeRabbit: Automated code review and PR feedback Corridor: Real-time security and threat detection Sourcery: Code review and generation assistance AI services - Infrastructure for building and scaling your own agents Braintrust: Evaluation and monitoring frameworks Kernel: Cloud browser infrastructure for agentic workloads Mixedbread: Multimodal AI search and retrieval across documents, code, media, and more Kubiks: Multi-step workflow orchestration and remediation Chatbase: Analytics and tuning for conversational agent Autonoma: AI-driven testing for your web and mobile apps to build, run, and analyze without writing code Browser Use: Natural language browser control for web automation Descope: No-code identity and authentication workflows fo

Open source Workflow Development Kit is now in public beta

October 23, 2025

Workflow Development Kit, a framework for building durable, long-running processes, is now in public beta. Workflow Development Kit brings durability, reliability, and observability to async JavaScript so you can build apps and AI agents that suspend, resume, and maintain state with ease. Turning functions into durable workflows is made simple by the "use workflow" directive: Key highlights include: Reliability by simply adding "use workflow" to make async functions durable. No manual wiring of queues, no schedulers, no YAML. Mark Steps to denote with "use step". Retries are automatic. Durability. Call sleep to pause without holding compute, then resume in place. Built-in observability. Traces, logs, and metrics for every run. Pause, replay, and time travel while debugging. No lock-in. Develop locally and deploy to Vercel or any other cloud. Learn more about Workflow or read the documentation. Read more

Vercel Python SDK is now available in beta

October 23, 2025

The Vercel Python SDK is now available in beta, bringing first-class Python support for many Vercel features such as Vercel Sandbox, Blob, and the Runtime Cache API. To get started, install the vercel package with pip install vercel. Vercel Python SDK lets you directly interact with Vercel primitives via python code like: Run untrusted code in isolated, ephemeral environments using Vercel Sandbox: Interact with Vercel’s Blob storage API: And store and retrieve data across Functions, Routing Middleware, and Builds within the same region using the Runtime Cache API: Get started with pip install vercel. Read more

Vercel Agent Investigations now in Public Beta

October 23, 2025

Vercel Agent can now run AI investigations on anomaly alerts to help teams diagnose and resolve incidents faster. AI investigations streamline incident response, improve production stability, and reduce alert fatigue to accelerate your team's shipping velocity. When an anomaly alert detects suspicious activity, such as unexpected spikes in usage or errors, Vercel Agent can investigate the issue, identify the likely root cause, analyze the impact, and suggest next steps for remediation. For greater control, you can also manually trigger an AI investigation directly from the anomaly alert details page. Vercel Agent investigations are now in public beta for Pro and Enterprise teams with Observability Plus. Pricing is usage-based, and teams can receive news: [00 Vercel Agent credits to get started. Try it out or learn more about Vercel Agent investigations. Read more

You can just ship agents

October 23, 2025

Building agents should feel like shaping an idea rather than fighting a maze of code or infrastructure. And we've seen this story before. A decade ago, the web moved from hand‑rolled routing and homegrown build scripts to opinionated frameworks and a platform that understood what developers were trying to do. Velocity went up, quality followed, and a generation of products appeared as if overnight. AI is following the same arc, but the stakes and surface area are larger because what you build is no longer a set of pages. It is a system that intelligently reasons, plans, and acts. Built on the foundations of Framework-defined Infrastructure, Vercel AI Cloud provides the tooling, infrastructure primitives, developer experience, and platform to bypass the complexity. You focus entirely on what you're building, with confidence in what's powering it under the hood. Read more

AI agents and services on the Vercel Marketplace

October 23, 2025

Agents and agentic AI give developers new ways to move faster and build better. They create connected, autonomous systems that continuously improve applications and raise the bar for speed and quality. But typically, integrating AI services means managing separate dashboards, billing systems, and authentication flows for each tool. A team using three different AI services might waste hours wiring up each integration before writing a single line of application code. Today, we're introducing the AI agents and services category to the Vercel marketplace. You can now add AI-powered workflows to your projects through native Vercel integrations with unified billing, observability, and installation flows built into the platform. Read more

Built-in durability: Introducing Workflow Development Kit

October 23, 2025

Building reliable software shouldn't require mastering distributed systems. Yet for developers building AI agents or data pipelines, making async functions reliable typically requires message queues, retry logic, and persistence layers. Adding that infrastructure often takes longer than writing the actual business logic. The Workflow Development Kit (WDK) is an open source TypeScript framework that makes durability a language-level concept. It runs on any framework, platform, and runtime. Functions can pause for minutes or months, survive deployments and crashes, and resume exactly where they stopped. Read more

Faster builds with Turbo build machines

October 22, 2025

Turbo build machines are now available for all paid plans, offering our fastest build performance yet with 30 vCPUs and 60GB of memory. Turbo machines are ideal for Turbopack builds, and large monorepos that run tasks in parallel, accelerating static generation and dependency resolution. Enable Turbo build machines per project, with usage-based pricing. Learn more in the documentation. Read more

Update regarding Vercel service disruption on October 20, 2025

October 21, 2025

At Vercel, our philosophy is to take ownership for, not blame, our vendors. Customers use our services to gain velocity, reliability, and ship wonderful products. Whether we picked A or B as one of the components of our “circuit design” is entirely our responsibility. Vercel is fully accountable for this incident, even if it's now public that it was triggered by the unexpected outage of AWS us-east-1 (called iad1 region). Vercel uses AWS infrastructure primitives, is part of the AWS marketplace, offers secure connectivity to AWS services, and shares a long history with AWS of pioneering serverless computing. To our customers, Vercel is unequivocally responsible for this outage. Our goal is to simplify the cloud and offer its best version. Through framework-defined infrastructure, we help developers focus on the application layer by deploying global infrastructure resources that are highly optimized. We operate our Compute, CDN, and Firewall services across 19 AWS regions, terminating and securing traffic in 95 cities and 130+ global points of presence. Yesterday, we fell short of this promise. While a significant amount of traffic was still served, and we shielded customers from th

Dynamically extend timeout of an active Sandbox

October 21, 2025

You can now extend the duration of a running Vercel Sandbox using the new extendTimeout method. This lets long-running sandboxes stay active beyond their initial timeout, making it easier to support workflows like chained agentic tasks or multi-step code generation that take longer than expected. You can extend the timeout multiple times until the maximum runtime for your plan is reached. Pro and Enterprise plans support up to 5 hours, with the Hobby plan supporting up to 45 minutes. Get started with Sandbox now and learn more in the docs. Read more

Preview links between microfrontends projects now serve all paths

October 21, 2025

Teams using microfrontends can now visit all routes from any domain in the microfrontends group, enabling teams to test their full site experience without broken links or missing pages. Previously, the microfrontend group's root domain would be the only one to serve the paths hosted by child microfrontends. Now, preview links between all microfrontends projects automatically serve all routes in the group. With this new feature: Preview links on child microfrontends now route paths to other microfrontends in the group, eliminating 404s. Deployments built from the same commit or branch automatically link to each other, making it easier to test changes in monorepos. Fallback routing ensures that requests to microfrontends not built on the same branch are still resolved. This feature is enabled by default for all new microfrontends, and will be rolling out slowly for existing teams. Learn more or get started with microfrontends today. Read more

Zero-configuration support for NestJS

October 17, 2025

Vercel now supports NestJS applications, a popular framework for building efficient, scalable Node.js server-side applications, with zero-configuration. Backends on Vercel use Fluid compute with Active CPU pricing by default. This means your NestJS app will automatically scale up and down based on traffic, and you only pay for what you use. Deploy NestJS on Vercel or visit the NestJS on Vercel documentation Read more

Braintrust joins the Vercel Marketplace

October 16, 2025

Braintrust is now available on the Vercel Marketplace, bringing AI evaluation and observability directly into the Vercel workflow. With this new integration, developers can automatically stream traces and evaluation data from Vercel to Braintrust with just a few clicks, gaining full visibility into model quality and user experience in real time. With Braintrust on Vercel Marketplace, you can: Ship agents and AI features with built-in evaluation and observability Run evals and monitor model quality in production Benchmark and compare performance across LLMs Explore the template to deploy the example today, with easy setup and unified billing. Read more

Talha Tariq joins Vercel as CTO of Security

October 15, 2025

As AI reshapes how software is built and deployed, the surface area for attacks is growing rapidly. Developers are shipping faster than ever, and we’re seeing new code paths, new threat models, and new vulnerabilities. That’s why I’m excited to share that Talha Tariq is joining Vercel as our CTO of Security. Talha brings deep expertise in security at scale, having served as CISO & CIO at HashiCorp for seven years before becoming CTO (Security) at IBM following its acquisition. There, he oversaw security across all IBM divisions including software, AI, and post-quantum cryptography. Read more

Just another (Black) Friday

October 15, 2025

For teams on Vercel, Black Friday is just another Friday. The scale changes, but your storefronts and apps stay fast, reliable, and ready for spikes in traffic. Many of the optimizations required for peak traffic are already built into the platform. Rendering happens at the edge, caching works automatically, and protection layers are on by default. What’s left for teams is refinement: confirming observability is set up, tightening security rules, and reviewing the dashboards that matter most. Last year, Vercel created a live Black Friday Cyber Monday dashboard that showcased our scale in real-time, showing the spikes. Overall, from Friday to Thursday, Vercel served 86,702,974,965 requests across its network, reaching a peak of 1,937,097 requests per second. Helly Hansen, a major technical apparel brand, entered the weekend with this confidence. Before the event, they moved from client-heavy rendering to Vercel’s CDN and saw: Read more

Introducing Trace Drains on the Vercel Marketplace

October 15, 2025

You can now use Vercel Drains to send traces and logs from your projects to your preferred Marketplace observability providers with native integrations, including Braintrust, Dash0, Statsig, and Kubiks with more providers coming soon. This integration allows developers to stream traces and evaluation data from Vercel directly into these providers for observability, debugging, and performance monitoring. The Trace Drain API extends Vercel’s observability surface to the Marketplace ecosystem, allowing providers to: Deliver rich visibility into performance and debugging data Integrate natively with logging, and analytics tools through Vercel Build tighter feedback loops between deployments and infrastructure insights Offer customers a fully connected experience without manual setup This update gives teams more flexibility to use their preferred observability tools while maintaining a single, unified developer experience inside Vercel. Try it out or learn more about this update available to Pro and ENT customers. Read more

Claude Haiku 4.5 now available in Vercel AI Gateway

October 15, 2025

You can now access Anthropic's latest model, Claude Haiku 4.5 using Vercel's AI Gateway with no other provider accounts required. Haiku 4.5 matches Sonnet 4's performance on coding, computer use, and agent tasks at substantially lower cost and faster speeds. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, set the model to anthropic/claude-haiku-4.5: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance and reliability to Claude Haiku 4.5, AI Gateway leverages multiple model providers under the hood, including Anthropic, Bedrock and Vertex AI. Learn more about AI Gateway, view the AI Gateway model leaderboard or try it in our model playground. Read more

Agents at work, a partnership with Salesforce and Slack

October 15, 2025

Every generation of software moves interfaces closer to where people think and work. Terminals gave way to GUIs. GUIs gave way to browsers. And now, the interface is language itself. Conversation has become the most natural way to build, explore, and decide. At the center of this shift is a new pattern: the AI agent. Today, software doesn’t have to wait for clicks or configuration, but understands user intent, reason about it, and takes action. The question for enterprises isn’t if they’ll adopt agents, but where those agents will live. Our answer: where work already happens. That’s why Vercel and Salesforce are partnering to help teams build, ship, and scale AI agents across the Salesforce ecosystem, starting with Slack. Together, we’re bringing the intelligence and flexibility of the Vercel AI Cloud to the places teams collaborate every day. Read more

Running Next.js inside ChatGPT: A deep dive into native app integration

October 15, 2025

When OpenAI announced the Apps SDK with Model Context Protocol (MCP) support, it opened the door to embedding web applications directly into ChatGPT. But there's a significant difference between serving static HTML in an iframe and running a full Next.js application with client-side navigation, React Server Components, and dynamic routing. This is the story of how we bridged that gap. We created a Next.js app that runs natively inside ChatGPT's triple-iframe architecture, complete with navigation and all the modern features you'd expect from a Next.js application. Read more

Commits to the same branch now build with no queues

October 14, 2025

Vercel now builds multiple commits to the same branch at the same time when On-Demand Concurrent Builds is enabled. Previously, a new commit would wait for the previous build on that branch to finish before starting. This update eliminates that queue, allowing commits to start building as soon as they arrive. Visit the On-demand concurrent builds documentation to learn more. Read more

Anomaly alerts now in public beta

October 14, 2025

Teams using Observability Plus can now receive alerts when anomalies are detected in their applications to help quickly identify, investigate and resolve unexpected behavior. Alerts help monitor your app in real-time by surfacing unexpected changes in usage or error patterns: Usage anomalies: unusual patterns in your application metrics, such as edge requests or function duration. Error anomalies: abnormal error patterns, such as sudden spikes in 5XX responses on a specific route View alerts directly in your dashboard, or subscribe via email, Slack, or webhooks to get notified wherever your team works. Alerts are available public beta for Pro and Enterprise customers with Observability Plus. Try it out or learn more about Alerts. Read more

Zero-configuration Flask backends

October 10, 2025

Flask, one of the most popular Python web application frameworks, can now be deployed instantly on Vercel with no configuration changes needed. Vercel's framework-defined infrastructure now recognizes and deeply understands Flask applications. This update removes the need for redirects in vercel.json or using the /api folder. Backends on Vercel use Fluid compute with Active CPU pricing by default. This means your Flask app will automatically scale up and down based on traffic, and you only pay for time where your code is actively using CPU. Deploy Flask on Vercel or visit the Flask on Vercel documentation. Read more

Expanded Role-Based Access Control (RBAC) for Enterprise teams

October 10, 2025

Vercel’s Role-Based Access Control (RBAC) system now supports multiple roles per user and introduces extended permissions for finer-grained access control across Enterprise teams. What’s new: Multi-role support: Assign multiple roles to a single user within Enterprise teams. Security role: A new team role dedicated to managing security and compliance settings. Extended permissions: Add granular capabilities that layer on top of team and project roles for precise control. Access groups integration: Access Groups now support team roles and extended permissions in Directory Sync mappings. The new extended permissions include: Create Project: Create new projects. Full Production Deployment: Deploy, rollback, and promote to production. Usage Viewer: View usage, prices, and invoices (read-only). Integration Manager: Install and manage integrations and storage. Environment Manager: Create and manage project environments. Environment Variable Manager: Create and manage environment variables. Extended permissions apply when paired with a compatible team role. Learn more in the Role-Based Access Control documentation. Read more

Block Vercel deployment promotions with GitHub Actions

October 9, 2025

You can now block a deployment from being promoted to production until selected GitHub Actions complete successfully. On Vercel, every deployment starts in a preview environment, this feature ensures that only verified builds that pass tests or other automated checks are released to production. Deployment Checks are available for all projects connected to GitHub repositories. Configure them in your project settings or learn more in the docs. Read more

Server rendering benchmarks: Fluid Compute and Cloudflare Workers

October 9, 2025

Independent developer Theo Browne recently published comprehensive benchmarks comparing server-side rendering performance between Fluid compute and Cloudflare Workers. The tests measured 100 iterations across Next.js, React, SvelteKit, and other frameworks. The results showed that for compute-bound tasks, Fluid compute performed 1.2 to 5 times faster than Cloudflare Workers, with more consistent response times. Read more

ChatGPT apps support on Vercel

October 9, 2025

You can now build and deploy ChatGPT apps directly on Vercel, with full support for modern web frameworks. ChatGPT apps let you integrate custom UI components and functionality within ChatGPT, deployed and served by Vercel. Frameworks like Next.js can now power these experiences using the Model Context Protocol (MCP), running natively inside the OpenAI sandbox rather than in a nested iframe. Check out our Next.js template. Build your ChatGPT apps with: Next.js features like server-side rendering (SSR) and React Server Components Vercel platform capabilities such as preview deployments, instant rollback, and a seamless dev-to-production pipeline Get started by building and deploying ChatGPT apps on Vercel using Next.js, Apps SDK, and mcp-handler. Read more

New Domains Registrar API for domain search, pricing, purchase, and management

October 8, 2025

You can now programmatically search, price, buy, renew, and transfer domains with Vercel’s new Domains Registrar API, complementing the new in-product Domains experience. The API provides endpoints for: Catalog & pricing: list supported TLDs; get TLD and per-domain pricing. Availability: check single or bulk availability. Orders & purchases: buy domains (including bulk) and fetch order status by ID. Transfers: retrieve auth codes, transfer in, and track transfer status. Management: renew, toggle auto-renewal, update nameservers, and fetch TLD-specific contact schemas. Explore the API docs. Read more

Anomaly alerts now available via email

October 7, 2025

Enterprise customers with Observability Plus can now receive anomaly alerts by email or in-app notifications, in addition to existing delivery options: webhooks, Slack, and the dedicated alerts dashboard. Currently, two types of anomaly alerts are available: Usage anomalies: Detects unusual spikes in key billable metrics Error anomalies: Detects sudden increases in 5XX responses on a specific route or path Anomaly alerts are available in limited beta for Enterprise customers with Observability Plus. Try it out or learn more about Alerts. Read more

Improved CLI experience when linking and creating environment variables​​​​‌‍​‍​‍‌‍ ‌​‍‌‍‍‌‌‍‌‌‍‍‌‌‍‍​‍​‍​‍‍​‍​‍‌‍​‌‍ ‌‍‍‌‌​‌‍‌‌‌‍‍‌‌​‌‍‌‍‌‌‌‌‍​​‍‍‌‍​‌‍ ‌‍‌​‍​‍​‍​​‍​‍‌‍‍​‌​‍‌‍‌‌‌‍‌‍​‍​‍​‍‍​‍​‍‌‍‍​‌‌​‌‌​‌​​‌​​‍‍​‍ ​‍

October 5, 2025

Here are some of the key improvements introduced in version 50.0.0: After successfully linking a project, the CLI now prompts you to pull your Project’s Environment Variables to keep your local setup aligned with your deployed configuration. Input for new Environment Variables is now masked during interactive entry When connecting to an existing project with link, the CLI now shows an interactive selector if you have fewer than 100 Projects. Fixed an issue where vc link --repo would incorrectly prefix project names. Commands that support the ls argument now have standardized behavior. Extra or unexpected arguments will consistently produce a clear error and exit early, ensuring predictable and reliable results across all ls commands. This change may require updates to scripts that depended on the previous behavior. Read more

Python package manager uv is now available for builds with zero configuration

October 3, 2025

Vercel now uses uv, a fast Python package manager written in Rust, as the default package manager during the installation step for all Python builds. This change makes builds 30-65% faster and adds support for more dependency formats. In addition to requirements.txt or Pipfile, projects can now declare dependencies with a uv.lock or pyproject.toml file. Learn more about the Python runtime on Vercel. Read more

Invalidate the CDN cache by tag

October 3, 2025

You can now invalidate CDN cache contents by tag. This marks all cached content associated with the tag as stale. The next request serves stale content instantly while revalidation happens in the background, with no latency impact for users. There are several ways to invalidate content: Dashboard Settings Vercel CLI Function API REST API In addition to invalidating by tag if the origin content changes, you can also delete by tag if the origin content is gone. However, deleting the cache can increase latency while new content is generated or cause downtime if your origin is unresponsive, so use with caution. Available on all plans. Learn more about cache invalidation. Read more

Static IPs are now available for more secure connectivity

October 2, 2025

Pro and Enterprise teams can now use Static IPs to securely connect to external services like databases that require IP allowlisting. Traffic from builds and functions routes through consistent, shared static IPs. To enable Static IPs, you can access Connectivity > Static IPs from your project or team settings. Static IPs are offered in addition to Secure Compute, which remains available for teams that need a fully dedicated VPC model. Note, Secure Compute has also moved within the Connectivity settings of your projects and teams. This is part of our move to bring more enterprise features self-serve. Read the docs or enable Static IPs here. Read more

Faster time-to-start for v0 builds

October 2, 2025

Publishing v0 apps is now 1.1s faster on average. We reduced the time it takes to send source files during deployment creation, improving the overall deployment pipeline and shortening feedback loops for developers. Deploy today on v0.app. Read more

Deployment-level configuration for Fluid compute

October 2, 2025

You can now configure Fluid compute on a per-deployment basis. By setting "fluid": true in your vercel.json, Fluid compute will be activated for that specific deployment. You can also enable or disable Fluid regardless of project level settings. This allows teams to selectively test and adopt Fluid compute without changing the global project settings. Read more in our documentation. Read more

Towards the AI Cloud: Our Series F

September 30, 2025

Today, Vercel announced an important milestone: a Series F funding round valuing our company at $9.3 billion. The $300M investment is co-led by longtime partners at Accel and new investors at GIC, alongside other incredible supporters. We're also launching a ~$300M tender offer for certain early investors, employees, and former employees. To all the customers, investors, and Vercelians who have been on this journey with us: thank you. Read more

Stripe is now available in beta on the Vercel Marketplace

September 30, 2025

Stripe is now available in beta on the Vercel Marketplace as a new payment provider. You can now provision a fully functional Stripe claimable sandbox directly from Vercel with no setup required. When ready, link it to a Stripe account and soon promote it to production. This makes it easy for teams to move from prototype to production for use cases like: Ecommerce storefronts: Test complete checkout flows before launch. SaaS billing: Validate subscriptions, usage-based pricing, and invoicing. Demos and templates: Share preconfigured environments for testing or client demos. Developer onboarding: Give teams instant access to ready-to-use Stripe sandboxes. Get started today with this example to build your first online simple store using Vercel and Stripe. Read more

View & query bot verification data in Vercel Observability

September 30, 2025

Vercel inspects every request to identify bot traffic. For requests claiming to come from a verified source, Vercel cross-checks against its directory of verified bots and validates them against strict verification criteria. We've added three new dimensions to the query builder when analyzing Edge Requests to help you understand bot activity to your projects: Bot name: Identify specific bots Bot category: Group bots by type Bot verified: Distinguish between verified, spoofed, and unverifiable bots Additionally, the Edge Requests dashboard in Observability now displays verification badges next to bot names. All users can view bot verification badge while Observability Plus subscribers can query this data at no extra cost. Try it out or learn more about Observability and Observability Plus. Read more

Collaborating with Anthropic on Claude Sonnet 4.5 to power intelligent coding agents

September 29, 2025

Claude Sonnet 4.5 is now available on Vercel AI Gateway with full support in AI SDK. We’ve been testing the model in v0, across our Next.js build pipelines, and inside our new Coding Agent Platform template. The model shows improvements in design sensibility and code quality, with measurable gains when building and linting Next.js applications. Claude Sonnet 4.5 builds on Anthropic's strengths in reasoning and coding. When paired with the Vercel AI Cloud, it powers a new class of developer workflows where AI can plan, execute, and ship changes safely inside your repositories. Read more

Node.js Vercel Functions now support per-path request cancellation

September 26, 2025

Vercel Functions using Node.js can now detect when a request is cancelled and stop execution before completion. This is configurable on a per-path basis, and includes actions like navigating away, closing a tab, or hitting stop on an AI chat to terminate compute processing early. This reduces unnecessary compute, token generation, and sending data the user would never see. To enable cancellation, add "supportsCancellation": true to your vercel.json configuration. You can apply it to specific paths or all functions: Once enabled, you can listen for cancellation using Request.signal.aborted or the abort event: If you’re using the AI SDK, forward the abortSignal to your stream: Learn more about cancelling Function requests. Read more

Preventing the stampede: Request collapsing in the Vercel CDN

September 25, 2025

When you deploy a Next.js app with Incremental Static Regeneration (ISR), pages get regenerated on-demand after their cache expires. ISR lets you get the performance benefits of static generation while keeping your content fresh. But there's a problem. When many users request the same ISR route at once and the cache is expired, each request can trigger its own function invocation. This is called a "cache stampede." It wastes compute, overloads your backend, and can cause downtime. The Vercel CDN now prevents this with request collapsing. When multiple requests hit the same uncached path, only one request per region invokes a function. The rest wait and get the cached response. Vercel automatically infers cacheability for each request through framework-defined infrastructure, configuring our globally distributed router. No manual configuration needed. Read more

Vercel Domains overhauled with instant search and at-cost pricing

September 25, 2025

We’ve rebuilt Vercel Domains end to end, making it faster, simpler, and more affordable to find and buy the right domain for your project. Search without login: Look up domains instantly, even when you’re not signed in. At-cost pricing: Domains are offered at registrar cost, with savings up to 50% on popular TLDs. Transparent results: Availability and pricing surface instantly, with no upsells or unnecessary add-ons. Fastest search on the web: Real-time, streaming results show availability and premium status instantly. Expanded TLD coverage: Support for more registries so every project can find the right home. Bulk checkout: Purchase multiple domains in a single streamlined transaction. This update makes Vercel Domains the fastest way to claim a name and get to production. As part of the overhaul, we’ve partnered with name.com as our upstream registrar, delivering better pricing and reliability. An upcoming blog will share how we built this speed using structured concurrency, layered caching, Bloom filters, and partitioned batching. Try it now at vercel.com/domains. Read more

Zero-configuration FastAPI backends

September 25, 2025

FastAPI, a modern, high-performance, web framework for building APIs with Python, is now supported with zero-configuration. Vercel's framework-defined infrastructure now recognizes and deeply understands FastAPI applications. This update removes the need for redirects in vercel.json or using the /api folder. Backends on Vercel use Fluid compute with Active CPU pricing by default. This means your FastAPI app will automatically scale up and down based on traffic, and you only pay for what you use. Deploy FastAPI on Vercel or visit the FastAPI on Vercel documentation. Read more

Request collapsing for ISR cache misses

September 25, 2025

The Vercel CDN now prevents cache stampedes through request collapsing on an expired Incremental Static Regeneration (ISR) page into a single function invocation per region. Without collapsing, simultaneous requests each trigger regeneration, wasting compute and overloading backends. With collapsing, one request regenerates the page while others wait and return the cached result. This improves reliability, reduces backend load, and saves significant compute at scale. The feature is applied automatically for cacheable routes. Cacheability is inferred from framework metadata, so no configuration is required. Implementation details are available in the Preventing the stampede: Request collapsing in the Vercel CDN blog post. Read more

Query data on external API requests in Vercel Observability

September 24, 2025

Observability Plus now supports querying and visualizing external API requests. Observability Plus's query builder allows customers to explore their Vercel data and visualize traffic, performance, and other key metrics. You can now author custom queries on request counts or time to first byte (TTFB) for external API calls, such as fetch requests to an AI provider. TTFB queries include breakdowns by average, min, max, p75, p90, p95, and p99. You can also filter or group results by request hostname to focus on specific APIs. The query builder is available to Pro and Enterprise teams using Observability Plus. Learn more about Observability and Observability Plus. Read more

Claimed deployments now include third-party resources

September 23, 2025

AI platforms, coding tools, and workflow apps can now create projects on Vercel that users can later claim as their own, transferring deployment ownership together with any resources by third-party providers. How it works: Instant deployment: Any third-party can use the Vercel API to create a project, deploy an application, and attach a resource store (such as a database). Claim and transfer: When a user claims the Vercel deployment, the attached resources automatically move with it. Full ownership of the complete deployment is handed off to the user. This is available today through Prisma, the first Vercel Marketplace provider to support instant deployment. Prisma customers can now spin up a database and a Vercel-hosted app together as a single, bundled stack. We’re expanding this flow to more Marketplace providers so they can pair their products, such as authentication, observability, and workflow services, with Vercel deployments through one-click claiming. Check out our Claim Deployments live demo and documentation to learn more. Read more

BotID uncovers hidden SEO poisoning

September 22, 2025

Your traffic is spiking and you spot suspicious bot activity in your logs. You deploy BotID expecting to find malicious scrapers, but the results show verified Google bots. Normal crawlers doing their job. But then you notice what they're actually searching for on your site. Queries that have nothing to do with your business. What do you do? This exact scenario recently played out at one of the largest financial institutions in the world. What they discovered was a years-old SEO attack still generating suspicious traffic patterns. Read more

Anomaly alerts now include error spikes

September 22, 2025

Enterprise customers with Observability Plus can now be alerted when error events deviate from normal behavior, helping teams catch issues earlier. The system automatically detects and groups abnormal error patterns, such as sudden spikes in 5XX responses on a specific route. Alert detail pages include relevant log lines, making it easier to investigate and resolve the underlying cause. Error anomaly detection, joining anomaly alerts for unusual app metric usage, is available in limited beta for Enterprise customers with Observability Plus Try it out or learn more about Alerts. If you have feedback or questions, drop them in the Vercel Community thread. Read more

Filter deployments by author

September 22, 2025

You can now filter deployments in the dashboard by author, using username, email, or Git username. Filters persist in the URL, making it easy to share filtered views with your team. See now in your project's deployments. Read more

How we made global routing faster with Bloom filters

September 19, 2025

Recently, we shipped an optimization to our global routing service that reduced its memory usage by 15%, improved time-to-first-byte (TTFB) from the 75th percentile and above by 10%, and significantly improved routing speeds for websites with many static paths. A small number of websites, with hundreds of thousands of static paths, were creating a bottleneck that slowed down our entire routing service. By replacing a slow JSON parsing operation with a Bloom filter, we brought path lookup latency down to nearly zero and improved performance for everyone. Read more

Observability Plus replacing legacy Monitoring

September 19, 2025

Observability Plus will be replacing the legacy Monitoring subscription for authoring custom queries on Vercel data. With Observability Plus, you have access to an expanded data set and more options for visualizing your Vercel data than was possible with legacy Monitoring. Observability Plus also allows you to save queries and visualizations to Notebooks to share insights and collaborate with team members. Monitoring will be sunset for Pro customers at the end date of their billing cycle in November. For Pro customers still subscribed to the legacy Monitoring SKU, we recommend subscribing to Observability Plus before their November billing cycle to continue to access custom queries on their Vercel data. Pro customers who have already subscribed to Observability Plus do not need to take any action. Learn more about authoring custom queries with Observability Plus. Read more

What you need to know about vibe coding

September 18, 2025

In February 2025, Andrej Karpathy introduced the term vibe coding: a new way of coding with AI, “[where] you fully give in to the vibes, embrace exponentials, and forget that the code even exists.” Just months later, vibe coding is completely reshaping how developers and non-developers work. Over 90% of U.S. developers use AI coding tools, adoption is accelerating for other roles, and English has become the fastest growing programming language in the world. We explore this shift in detail in our new State of Vibe Coding. Here are a few of the key takeaways. Read more

Scale to one: How Fluid solves cold starts

September 18, 2025

Cold starts have long been the Achilles’ heel of traditional serverless. It’s not just the delay itself, but when the delay happens. Cold starts happen when someone new discovers your app, when traffic is just starting to pick up, or during those critical first interactions that shape whether people stick around or convert. Traditional serverless platforms shut down inactive instances after a few minutes to save costs. But then when traffic returns, users are met with slow load times while new instances spin up. This is where developers would normally have to make a choice. Save money at the expense of unpredictable performance, or pay for always-on servers that increase costs and slow down scalability. But what if you didn't have to choose? That’s why we built a better way. Powered by Fluid compute, Vercel delivers zero cold starts for 99.37% of all requests. Fewer than one request in a hundred will ever experience a cold start. If they do happen, they are faster and shorter-lived than on a traditional serverless platform. Through a combination of platform-level optimizations, we've made cold starts a solved problem in practice. What follows is how that’s possible and why it works

AI code reviews by Vercel Agent now in Public Beta

September 18, 2025

Vercel Agent can now conduct code reviews, with validated suggestions that address issues across correctness, security, and performance. These reviews are fully codebase-aware, looking beyond the diff to any relevant files. Proposed patches are generated and validated in Vercel Sandboxes before they ever reach your PR. Key features also include: Optimizations for frameworks like Next.js, React, Nuxt, and Svelte, with support for TypeScript, Python, Go, and more High-signal, inline comments for human review including diffs, analysis, and reproduction steps for transparency In-dashboard Observability for metrics like files read, review time, cost, and more Configuration options to review all, public, or private repositories, and skip draft PRs Available in public beta for all Pro and Enterprise teams, pricing is fully usage based with a news: [00 Vercel Agent credit. Try it for free in the new Agent dashboard, read more in the docs, or provide feedback in Vercel Community. Read more

Addressing security and quality issues with MCP tools in AI Agent

September 17, 2025

Model Context Protocol (MCP) is emerging as a standard protocol for federating tool calls between agents. Enterprises are starting to adopt MCP as a type of microservice architecture for teams to reuse each other's tools across different AI applications. But there are real risks with using MCP tools in production agents. Tool names, descriptions, and argument schemas become part of your agent's prompt and can change unexpectedly without warning. This can lead to security, cost, and quality issues even when the upstream MCP server has not been compromised or is not intentionally malicious. We built mcp-to-ai-sdk to reduce these issues. It is a CLI that generates static AI SDK tool definitions from any MCP server. Definitions become part of your codebase, so they only change when you explicitly update them. Read more

Builds now start up to 30% faster

September 16, 2025

The build cache stores files from previous builds to speed up future ones. We've improved its performance by downloading parts of the cache in parallel using a worker pool. This decreased the build initialization time by 30% on average, reducing build times by up to 7 seconds for all plans. This is enabled automatically for all new builds and adds to the build initialization improvements previously launched. Learn more about builds on Vercel. Read more

AI agents at scale: Rox’s Vercel-powered revenue operating system

September 16, 2025

Rox is building the next-generation revenue operating system. By deploying intelligent AI agents that can research, prospect, and engage on behalf of sellers, Rox helps enterprises manage and grow revenue faster. From day one, Rox has built their applications on Vercel. With Vercel's infrastructure powering their web applications, Rox ships faster, scales globally, and delivers consistently fast experiences to every customer. Read more

Shai-Halud Supply Chain Campaign — Expanded Impact & Vercel Response

September 16, 2025

Summary The Shai-Halud supply chain campaign has escalated. What began with the Qix compromise affecting ~18 core npm packages (chalk, debug, ansi-styles, etc.) has since spread: Over 40 additional packages attacked via the Tinycolor “worm” vector. The CrowdStrike / crowdstrike-publisher namespace was also compromised, with multiple trojanized releases. The DuckDB maintainer account (duckdb_admin) published malicious versions matching the same wallet-drainer malware used in the Qix incidents. No Vercel customers were impacted in that DuckDB subset. Impact to Vercel Customers We identified a small set of 10 Vercel customer projects whose builds depended (directly or transitively) on the compromised package versions. Impacted customers have been notified and provided with project-level guidance. In the DuckDB incident, no Vercel customer build was affected. What We Did Action Status Blocklisted known compromised versions from the Tinycolor, CrowdStrike, Qix, and DuckDB-affected packages ✅ Completed Purged build caches for Vercel projects using those versions ✅ Completed for impacted projects Coordinated safe rebuilds with clean dependencies / pinned safe versions ✅ In progress / comp

Helly Hansen migrated to Vercel and drove 80% Black Friday growth

September 15, 2025

Founded in 1877, Helly Hansen is a global leader in technical apparel, but its digital experience wasn't living up to its legacy. Operating across 38 global markets with multiple brands (including HellyHansen.com, HHWorkwear.com, and Musto.com), the company was being held back by an outdated tech stack that slowed site speeds and frustrated customers. Through an incremental migration to Next.js and Vercel, Helly Hansen improved Core Web Vitals from red to green, increased developer agility, and delivered a record-breaking Black Friday Cyber Monday, building a foundation for future innovation. Read more

Introducing Vercel Drains: Complete observability data, anywhere

September 15, 2025

Vercel Log Drains are now Vercel Drains. Why? They’re not just for logs anymore, as you can now also export OpenTelemetry traces, Web Analytics events, and Speed Insights metrics. Drains give you a single way to stream observability data out of Vercel and into the systems your team already rely on. Read more

Updated defaults for deployment retention

September 15, 2025

Starting October 15, 2025, Vercel will update the default deployment retention policy for all projects currently using the legacy “unlimited” setting: Canceled Deployments - 30 days, with a maximum of 1 year. Errored Deployments - 3 months, with a maximum of 1 year. Pre-Production Deployments - 6 months, with a maximum of 3 years. Production Deployments - 1 year, with a maximum of 3 years. Projects with a custom deployment retention setting will not be affected. Additionally, before October 15, the "unlimited" option will become unavailable when modifying retention policies. Team owners can configure a default retention policy to be applied to any new projects created under the team on Teams > Security & Privacy > Deployment Retention Policy. This policy can also be easily applied to all existing projects. Note that your 10 most recent production deployments and any currently aliased deployment will never be deleted, regardless of age. Learn more about Deployment Retention. Read more

New Vercel CLI login flow

September 12, 2025

The vercel login command now uses the industry-standard OAuth 2.0 Device Flow, making authentication more secure and intuitive. You can sign in from any browser-capable device. When approving a login, be sure to verify the location, IP, and request time before granting access to your Vercel account. Email-based login (vercel login your@email.com) and the flags --github, --gitlab, --bitbucket, --oob, and team are deprecated. These methods will no longer be supported beginning February 26, 2026, except for the team method (SAML-based login), which remains supported until June 1, 2026. Note: Support had previously been extended from the original deprecation date of February 1, 2026 to June 1, 2026. To prioritize user security, we are moving the deprecation date forward: most of these methods will be removed on February 26, 2026, with the team method following on June 1, 2026. Upgrade today with npm i vercel@latest Learn more in the docs. Read more

Introducing x402-mcp: Open protocol payments for MCP tools

September 12, 2025

AI agents are improving at handling complex tasks, but a recurring limitation emerges when they need to access paid external services. The current model requires pre-registering with every API, managing keys, and maintaining separate billing relationships. This workflow breaks down if an agent needs to autonomously discover and interact with new services. x402 is an open protocol that addresses this by adding payment directly into HTTP requests. It uses the 402 Payment Required status code to let any API endpoint request payment without prior account setup. We built x402-mcp to integrate x402 payments with Model Context Protocol (MCP) servers and the Vercel AI SDK. Read more

Qwen3-Next models are now supported in Vercel AI Gateway

September 12, 2025

You can now access Qwen3 Next, two ultra-efficient models from QwenLM, designed to activate only 3B parameters, using Vercel's AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to alibaba/qwen3-next-80b-a3b-thinking: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway and access the model here. Read more

402-mcp enables x402 payments in MCP

September 12, 2025

Introducing x402-mcp, a library that integrates with the AI SDK to bring x402 paywalls to Model Context Protocol (MCP) servers to let agents discover and call pay for MCP tools easily and securely. With x402-mcp, you can define MCP servers with paidTools that require payment to run, enabling account-less, low-latency, anonymous payments directly in AI workflows. Payments confirm in ~100–200ms, with fees under $0.01 and support for minimums under $0.001. Getting started is easy, here's how you can define a paid tool: And integrating with AI SDK MCP Clients takes just one function to enable payments: Read more about x402 or try our full stack x402 AI Starter Kit. Read more

LongCat-Flash Chat model is now supported in Vercel AI Gateway

September 11, 2025

You can now access LongCat Flash Chat, a new model from Meituan focused on agentic tool use, using Vercel AI Gateway with no other provider accounts required. The model dynamically activates parameters, based on contextual demands. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to meituan/longcat-flash-chat: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway and access the model here. Read more

MongoDB Atlas is now available on the Vercel Marketplace

September 10, 2025

MongoDB Atlas is now available on the Vercel Marketplace. Developers can now provision a fully managed MongoDB database directly from your Vercel dashboard and connect it to your project without leaving the platform. Adding a database to your project typically means managing another account, working through connection setup, and coordinating billing across services. The Vercel Marketplace brings these tools into your existing workflow, so you can focus on building rather than configuring. Read more

ChatGPT can now integrate with Vercel MCP

September 10, 2025

You can now use ChatGPT with Vercel MCP, our official Model Context Protocol (MCP) server. For security, Vercel MCP currently supports AI clients that have been reviewed and approved by Vercel. Connectors within ChatGPT are available in beta to Pro and Plus accounts on the web. Follow the steps below to set up Vercel as a connector within ChatGPT: Enable developer mode: Go to Settings → Connectors → Advanced → Developer mode. Add Vercel MCP Open ChatGPT settings In the Connectors tab, click Create Name: Vercel MCP server URL: https://mcp.vercel.com. Authentication: OAuth Click Create You should now be able to select Vercel as a connector in Developer Mode chats. With Vercel MCP you can give agents access to protected deployments, analyze build logs, and more. Read more about using AI tools with Vercel MCP. Read more

MongoDB Atlas joins the Vercel Marketplace

September 10, 2025

You can now provision MongoDB Atlas directly from the Vercel Marketplace. Spin up a fully managed MongoDB Atlas database, connect it to you Vercel project, and start building without leaving the Vercel dashboard. This MongoDB Atlas native integration provides: A flexible document model for structured and unstructured data Built-in search, including vector and semantic search Horizontal scaling with replica sets and sharding Free, pre-provisioned, or serverless deployment options This integration removes the friction of switching between dashboards or managing complex setup, giving developers a fast, modern data layer to power web and AI applications on Vercel. Get started with MongoDB Atlas on the Vercel Marketplace, available to customers on all plans. Learn more in the blog post and deploy the MongoDB Atlas Forum template on Vercel. Read more

Vercel Sandbox maximum duration extended to 5 hours

September 10, 2025

Pro and Enterprise teams can now run Vercel Sandboxes for up to 5 hours (up from 45 minutes). This new max duration unlocks workloads that require longer runtimes, such as large-scale data processing, end-to-end testing pipelines, and long-lived agentic workflows. Get started with Sandbox now and learn more in the docs. Read more

A more flexible Pro plan for modern teams

September 9, 2025

We’re updating Vercel’s Pro plan to better align with how modern teams collaborate, how applications consume infrastructure, and how workloads are evolving with AI. Concretely, we’re making the following changes: Read more

HIPAA BAAs are now available to Pro teams

September 9, 2025

Pro teams can now enter into a Business Associate Agreement (BAA) to support HIPAA-compliant workloads on Vercel. The BAA is available self-serve through the dashboard with no Enterprise contract required. Vercel supports HIPAA compliance as a business associate by implementing technical and organizational safeguards, conducting annual audits, and offering breach notification in line with HIPAA requirements. Compliance is a shared responsibility between you and Vercel. Teams are responsible for configuring security features, managing access, and validating third-party services. This update makes it easier for healthcare-focused applications to meet regulatory obligations without upgrading to Enterprise. Frequently asked questions: Read more about other updates to Pro and switch to the new Pro pricing. Read more

No build queues: On-demand concurrent builds now on by default

September 9, 2025

Teams on the new Pro pricing model will now have on-demand concurrent builds enabled by default. This ensures builds across projects start immediately without waiting in a queue, except when multiple builds target the same Git branch. You can manage this setting at any time using the new bulk enable feature, even if your team is not yet on the new Pro pricing model. Learn more in the documentation read more about the updates to Pro, and switch to the new Pro pricing. Read more

Spend Management now enabled by default on Pro

September 9, 2025

Spend Management is now enabled for new Pro teams, and will be enabled by default for existing teams when they switch to the new pricing model. All Pro teams will have a budget set by default based on any previous usage, if any, and any teams with existing budgets will be unaffected. This can be changed at any time in spend management settings. Email alerts will be sent when nearing the spend threshold based on your on-demand spend. This can be changed at any time here. This new default ensures you receive proactive cost signals to manage your spend. Your deployments will continue without interruption unless a hard limit is manually configured. Read more about the updates to Pro and switch to the new Pro pricing. Read more

Free Viewer seats now available on Pro

September 9, 2025

Pro teams can now add unlimited Viewer seats at no additional charge so team members can collaborate more flexibly and cost-efficiently. Viewers can access project dashboards, deployments, analytics, and more, but can’t see sensitive data, deploy, or change production settings. This is ideal for any team members looking to collaborate via access to the dashboard and preview deployments. Previously, all seats on Pro were paid. Moving forward, you can add two main types of seats: Developer seats (Owner, Member): Team members that deploy, debug, and configure. These seats remain $20. Viewer seats: Team members that do not deploy. These seats are free. Viewers can easily request an upgrade from team owners directly in the dashboard. Read more about the updates to Pro and switch to the new Pro pricing. Read more

Included Pro usage is now credit-based

September 9, 2025

The Pro plan now includes $20 in monthly usage credit instead of fixed allocations across metrics like data transfer, compute, caching, and more. This plan update replaces static usage buckets with a more flexible system that adapts to your workload. In addition to the above, the new Pro pricing model includes: Free viewer seats Self-serve enterprise features like SAML SSO and HIPAA BAA Better Spend Management, enabled by default Read more about the updates to Pro and switch to the new Pro pricing. Read more

The second wave of MCP: Building for LLMs, not developers

September 9, 2025

When the MCP standard first launched, many teams rushed to ship something. Many servers ended up as thin wrappers around existing APIs with minimal changes. A quick way to say "we support MCP". At the time, this made sense. MCP was new, teams wanted to get something out quickly, and the obvious approach was mirroring existing API structures. Why reinvent when you could repackage? But the problem with this approach is LLMs don’t work like developers. They don’t reuse past code or keep long term state. Each conversation starts fresh. LLMs have to rediscover which tools exist, how to use them, and in what order. With low level API wrappers, this leads to repeated orchestration, inconsistent behavior, and wasted effort as LLMs repeatedly solve the same puzzles. MCP works best when tools handle complete user intentions rather than exposing individual API operations. One tool that deploys a project end-to-end works better than four tools that each handle a piece of the deployment pipeline. Read more

Streamdown 2.2 - animated streaming and better support for custom HTML

September 8, 2025

Streamdown 2.2 delivers animated per-word text streaming, improved custom HTML handling, and drop-in compatibility with ReactMarkdown - making it easier to adopt Streamdown in existing projects. Animated streaming By importing the Streamdown stylesheet and enabling the new animated prop, streaming content renders with smooth per-word text animation. This provides a polished experience for AI chat interfaces and other real-time text applications. Better custom HTML support The components prop now accepts custom HTML attributes by adding elements to allowedTags. The Remend engine has also been improved to strip incomplete HTML tags during streaming, preventing visual glitches from partial markup. ReactMarkdown compatibility Streamdown now supports the remaining ReactMarkdown props, making migration a one-line replacement. Projects using ReactMarkdown can swap to Streamdown without refactoring component configurations. This release also removes CommonJS builds, adds bundled-language aliases for common JavaScript, TypeScript, and shell labels, and includes numerous rendering and security fixes across tables, code blocks, LaTeX, and Mermaid diagrams. Learn more in the Streamdown docs. R

Package installation for v0 builds is now ~70% faster.

September 8, 2025

Average npm install time for v0 builds dropped from 5s to 1.5s by optimizing how dependencies are resolved and cached during build execution. This is in addition to a recent improvement to time-to-start for v0 builds, with more improvements in progress to further reduce installation and overall build time. Deploy today on v0.app. Read more

AI SDK and AI Gateway now integrated in GitHub Actions

September 8, 2025

You can now use the vercel/ai-action@v2 GitHub Action to access the AI SDK and AI Gateway, generating text or structured JSON directly in your workflows by specifying a prompt, model, and api-key. This integration enables new AI-powered use cases for GitHub Actions, such as summarizing what made it into a release, providing a light PR code review, moderating comments, or finding duplicate or relevant issues. Learn more and see examples in the Github Actions marketplace, view the source code, or read the walkthrough guide. Read more

Bulk enable on-demand concurrent builds across projects

September 8, 2025

Pro teams can now remove build queues across all projects with just one click by bulk enabling on-demand concurrent builds. On-demand concurrency scales build compute capacity dynamically, so all builds for a project start as soon as they are requested, except when multiple builds target the same Git branch. To get started, visit your Pro team's billing settings to: Enable or disable for all existing projects Search and pick specific projects where it should be active Learn more in the documentation. Read more

Vercel Functions now support graceful shutdown

September 8, 2025

Vercel Functions running on Node.js and Python runtimes now support graceful shutdown, giving you up to 500 milliseconds to run cleanup tasks before termination. When a function is terminated, such as during scale-down, the runtime receives a SIGTERM signal. You can now use this signal to run cleanup tasks like closing database connections or flushing external logs. Learn more about the SIGTERM signal. Read more

Critical npm supply chain attack response – September 8, 2025

September 8, 2025

On September 9, 2025, the campaign extended to DuckDB-related packages after the duckdb_admin account was breached. These releases contained the same wallet-drainer malware, confirming this was part of a coordinated effort targeting prominent npm maintainers. While Vercel customers were not impacted by the DuckDB incident, we continue to track activity across the npm ecosystem with our partners to ensure deployments on Vercel remain secure by default. Read more

Export traces, web analytics events, and speed insights datapoints to any destination

September 5, 2025

Users can export OpenTelemetry traces, Web Analytics events, and Speed Insights data points from Vercel to any third-party tool. We’ve expanded our Log Drains infrastructure, enabling users to stream more raw data out of Vercel and into external systems. With Vercel Drains, users can configure custom HTTP endpoints to receive data in multiple encodings — JSON, NDJSON, or Protobuf. Pro and Enterprise teams can export data to external systems at the same $0.50 per GB rate. Try it out or learn more about Vercel Drains. Read more

Zero-configuration Express backends

September 5, 2025

Express, a fast, unopinionated, minimalist web framework for Node.js, is now supported with zero-configuration. Vercel's framework-defined infrastructure now recognizes and deeply understands Express applications. This update removes the need for redirects in vercel.json or using the /api folder. Deploy Express on Vercel or visit the Express on Vercel documentation. Read more

Stress testing Biome's noFloatingPromises lint rule

September 4, 2025

Recently we partnered with the Biome team to strengthen their noFloatingPromises lint rule to catch more subtle edge cases. This rule prevents unhandled Promises, which can cause silent errors and unpredictable behavior. Once Biome had an early version ready, they asked if we could help stress test it with some test cases. At Vercel, we know good tests require creativity just as much as attention to detail. To ensure strong coverage, we wanted to stretch the rule to its limits and so we thought it would be fun to turn this into a friendly internal competition. Who could come up with the trickiest examples that would still break the updated lint rule? Part of the fun was learning together, but before we dive into the snippets, let’s revisit what makes a Promise “float”. Read more

Moonshot AI's Kimi K2 0905 model is now supported in Vercel AI Gateway

September 4, 2025

You can now access Kimi K2 0905, a new model from Moonshot AI focused on agentic coding with a 256K context window, using Vercel AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to moonshotai/kimi-k2-0905: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance to Kimi K2, AI Gateway leverages multiple model providers under the hood, including direct to Moonshot AI, Groq, and Fireworks AI. Learn more about AI Gateway. Read more

Open SDK strategy

September 3, 2025

At Vercel, our relationship with open source is foundational. We do not build open source software to make money. Rather, we’re building an enduring business that enables us to continue developing great open source software. We believe in improving the default quality of software for everyone, everywhere, whether they are Vercel customers or not. A rising tide lifts all boats. Read more

CVE-2025-57822

August 29, 2025

Summary A vulnerability affecting Next.js Middleware has been addressed. It impacted versions prior to v14.2.32 and v15.4.7, and involved a Server-Side Request Forgery (SSRF) risk introduced by misconfigured usage of the NextResponse.next() function within middleware. Applications that reflected a user's request headers in this function, rather than passing them through the request object, could unintentionally allow the server to issue requests to attacker-controlled destinations. A patch applied on August 25th, 2025 eliminated exposure for Vercel customers running the affected versions. Impact In affected configurations, an attacker could: Influence the destination of internal requests triggered by middleware routing logic Perform SSRF against internal infrastructure if user-controlled headers (e.g., Location) were forwarded or interpreted without validation Potentially access sensitive internal resources or services unintentionally exposed via internal redirect behavior This issue is exploitable in self-hosted deployments where developers use custom middleware logic and do not adhere to documented usage of NextResponse.next({ request }). It is not exploitable on Vercel infrastru

CVE-2025-55173

August 29, 2025

Summary A vulnerability affecting Next.js Image Optimization has been addressed. It impacted versions prior to v15.4.5 and v14.2.31, and involved a scenario where attacker-controlled external image servers could serve crafted responses that result in arbitrary file downloads with attacker-defined filenames and content. Your Vercel deployments are safe by default. A patch applied on July 29th, 2025 eliminated exposure for all Vercel-hosted customers. Self-hosted deployments should upgrade to v15.4.5 or v14.2.31 to remediate the issue. Impact Under certain configurations (images.domains or permissive images.remotePatterns), a malicious actor could: Trigger the download of a file from a Next.js app with attacker-controlled content and filename Exploit this behavior for phishing, drive-by downloads, or social engineering scenarios This issue requires that: The target app has external image domains or patterns configured The remote server is attacker-controlled or attacker-influenced A user is tricked into clicking a crafted URL Resolution The issue was resolved by updating the image optimizer logic to avoid falling back to the upstream’s Content-Type header when magic number detection

CVE-2025-57752

August 29, 2025

Summary A vulnerability affecting Next.js Image Optimization has been addressed. It impacted versions prior to v15.4.5 and v14.2.31, and involved a cache poisoning issue that caused sensitive image responses from API routes to be cached and subsequently served to unauthorized users. Vercel deployments were never impacted by this vulnerability. Impact When API routes are used to return image content that varies based on headers (e.g., Cookie, Authorization), and those images are passed through Next.js Image Optimization, the optimized image may be cached without including those request headers as part of the cache key. This can lead to: Unauthorized disclosure of user-specific or protected image content Cross-user leakage of conditional content via CDN or internal cache This issue arises without user interaction and requires no elevated privileges, only a prior authorized request to populate the cache. Resolution The issue was resolved by ensuring request headers aren’t forwarded to the request that is proxied to the image endpoint. This ensures that the image endpoint cannot be used to serve images that require authorization data and thus cannot be cached. Fix available in: Next.js

Preparing for the worst: Our core database failover test

August 28, 2025

Many engineering teams have disaster recovery plans. But unless those plans are regularly exercised on production workloads, they don’t mean much. Real resilience comes from verifying that systems remain stable under pressure. Not just in theory, but in practice. On July 24, 2025, we successfully performed a full production failover of our core control-plane database from Azure West US to East US 2 with zero customer impact. This was a test across all control-plane traffic: every API request, every background job, every deployment and build operation. Preview and development traffic routing was affected, though our production CDN traffic, served by a separate globally-replicated DynamoDB architecture, remained completely isolated and unaffected across our 19 regions. This operation was a deliberate, high-stakes exercise. We wanted to ensure that if the primary region became unavailable, our systems could continue functioning with minimal disruption. The result: a successful failover with zero customer downtime, no degraded performance in production, and no postmortem needed. Read more

s1ngularity: supply chain attack in Nx packages

August 27, 2025

Threat actors published modified versions of the Nx package and some of its supporting libraries to the npm registry with the goal of exfiltrating developer and service credentials. Builds on Vercel are safe from this vulnerability by default. Visit the GitHub advisory to check if your local or other CI environments are impacted. Summary A malicious version of the Nx package and some Nx ecosystem libraries were published to the npm registry using a stolen npm token, starting at 6:32 PM EDT on August 26, 2025. The compromised packages were removed from the npm registry by the Nx team, ending at 10:44 PM EDT on the same day. The affected packages contained a postinstall script that scanned the user's file system using an LLM to exfiltrate secrets and credentials when installing an affected package. Exfiltrated secrets were posted as an encoded string into a GitHub repo that the script would create in the victim's GitHub account. For more information, visit the advisory on GitHub from the Nx team. Impact for Vercel customers By default, Vercel customers are not impacted, and can only be affected by the compromised Nx packages if they took specific steps leveraging the build container'

Anomaly alerts now in limited beta for Enterprise customers

August 27, 2025

Enterprise customers can now receive alerts when anomalies are detected in their applications, in order to quickly identify and mitigate issues. Anomaly detection: Automatically identifies unusual patterns in your application metrics. Webhook integration: Subscribe to alerts and route them into your existing monitoring systems. Slack notifications: Get alerts delivered directly to your team channels. Alerts are available in limited beta for Enterprise customers with Observability Plus. Try it out or learn more about Alerts. Read more

Deploy Slack's Bolt.js to Vercel with @vercel/slack-bolt

August 27, 2025

We've published @vercel/slack-bolt, our official adapter for deploying Slack's Bolt for JavaScript to Vercel's AI Cloud. Bolt provides a type-safe library for responding to Slack webhook events. However, Slack's API requires a response within three seconds or users are faced with timeouts. This has made it hard to build Slack agents on traditional serverless platforms. Our adapter uses Fluid compute’s streaming and waitUntil to acknowledge responses within Slack’s deadline while your agent continues working in the background. This adapter works with any function or framework using the Web API Request object such as Hono, Nitro or Next.js. Get started with our Slack Agent Template today or visit the library on npm. Read more

SAML SSO is now available to Pro teams

August 26, 2025

SAML-based Single Sign-On (SAML SSO) is now available as an add-on to all Pro teams and can be configured directly in the dashboard. This includes support for major identity providers like Okta, Azure AD, and Google Workspace. Previously limited to Enterprise plans, SAML SSO on Pro enables secure, centralized access control without requiring a contract. Read more about other updates to Pro and switch to the new Pro pricing. Read more

30-day runtime log retention, now available in Observability Plus

August 26, 2025

Teams with Observability Plus now have 30 days of runtime log retention. These logs include detail about requests, Vercel Functions and Routing Middleware invocations, cache activity, and more. You can view, query, inspect, and share up to 14 consecutive days of log data at once. This extended retention is available at no additional cost for Pro and Enterprise plans with Observability Plus enabled. Try it out or learn more about Runtime Logs. Read more

Devin, Raycast, Windsurf, and Goose now supported on Vercel MCP

August 25, 2025

You can now use Devin, Raycast, Windsurf, and Goose with Vercel MCP, our official Model Context Protocol (MCP) server. For security, Vercel MCP currently supports AI clients that have been reviewed and approved by Vercel. Follow the steps below to get started with each client: Devin Navigate to Devin's Settings > MCP Marketplace Search for Vercel and select the MCP Click Install Raycast Run the Install Server command Enter the following details: Name: Vercel Transport: HTTP URL: https://mcp.vercel.com Click Install Windsurf Add the snippet below to your mcp_config.json file Goose Click here for a one-click installation of the Vercel MCP. With Vercel MCP you can give agents access to protected deployments, analyze build logs, and more. Read more about using AI tools with Vercel MCP. Read more

AI-powered prototyping with design systems

August 22, 2025

Prototyping with AI should feel fast, collaborative, and on brand. Most AI tools have cracked the "fast" and "collaborative" parts, but can struggle with feeling "on-brand". This disconnect usually stems from a lack of context. For v0 to produce output that looks and feels right, it needs to understand your components. That includes how things should look, how they should behave, how they work together, and all of the other nuances. Most design systems aren’t built to support that kind of reasoning. However, a design system built for AI enables you to generate brand-aware prototypes that look and feel production ready. Let's look at why giving v0 this context creates on-brand prototypes and how you can get started. Read more

Deploy xmcp servers with zero-configuration

August 22, 2025

Vercel now supports xmcp, a framework for building and shipping MCP servers with TypeScript, with zero-configuration. xmcp uses file-based routing to create tools for your MCP server. Once you've created a file for your tool, you can use a default export in a way that feels familiar to many other file-based routing frameworks. Below, we create a "greeting" tool. Learn more about deploying xmcp to Vercel in the documentation. Read more

AI Gateway: Production-ready reliability for your AI apps

August 21, 2025

Building an AI app can now take just minutes. With developer tools like the AI SDK, teams can build both AI frontends and backends that accept prompts and context, reason with an LLM, call actions, and stream back results. But going to production requires reliability and stability at scale. Teams that connect directly to a single LLM provider for inference create a fragile dependency: if that provider goes down or hits rate limits, so does the app. As AI workloads become mission-critical, the focus shifts from integration to reliability and consistent model access. Fortunately, there's a better way to run. AI Gateway, now generally available, ensures availability when a provider fails, avoiding low rate limits and providing consistent reliability for AI workloads. It's the same system that has powered v0.app for millions of users, now battle-tested, stable, and ready for production for our customers. Read more

AI Gateway is now generally available

August 21, 2025

AI Gateway is now generally available, providing a single unified API to access hundreds of AI models with transparent pricing and built-in observability. With sub-20ms latency routing across multiple inference providers, AI Gateway delivers: Transparent pricing with no markup on tokens (including Bring Your Own Keys) Automatic failover for higher availability High rate limits Detailed cost and usage analytics You can use AI Gateway with the AI SDK or through the OpenAI-compatible endpoint. With the AI SDK, it’s just a simple model string switch. Get started with a single API call: Read more about the announcement, learn more about AI Gateway, or get started now. Read more

Introducing Streamdown: Open source Markdown for AI streaming

August 21, 2025

Streamdown is a new open source, drop-in Markdown renderer built for AI streaming. It powers the AI Elements Response component, but can also be used standalone to give developers a fully composable, independently managed option with npm i streamdown. Streamdown is designed to handle unterminated chunks, interactive code blocks, math, and other cases that are unreliable with existing Markdown packages. It's available now, and ships with: Tailwind typography styles: Preconfigured classes for headings, lists, and code blocks GitHub Flavored Markdown: Tables, task lists, and other GFM features Interactive code blocks: Shiki highlighting with built-in copy button Math support: LaTeX expressions via remark-math and KaTeX Graceful chunk handling: Proper formatting for unterminated Markdown chunks Security hardening: Safe handling of untrusted content with restricted images and links You can get started with start with AI Elements: Or as a standalone package: Read the docs and upgrade your AI-powered streaming. Read more

Rethinking prototyping, requirements, and project delivery at Code and Theory

August 20, 2025

Code and Theory is a digital-first creative and technology agency that blends strategy, design, and engineering. With a team structure split evenly between creatives and engineers, the agency builds systems for global brands like Microsoft, Amazon, and NBC that span media, ecommerce, and enterprise tooling. With their focus on delivering expressive, scalable digital experiences, the team uses v0 to shorten the path from idea to working software. Read more

<script type="text/llms.txt">

August 20, 2025

How do you tell an AI agent what it needs to do when it hits a protected page? Most systems rely on external documentation or pre-configured knowledge, but there's a simpler approach. What if the instructions were right there in the HTML response? llms.txt is an emerging standard for making content such as docs available for direct consumption by AIs. We’re proposing a convention to include such content directly in HTML responses as . Read more

Agents can now access protected deployments via Vercel’s MCP server

August 19, 2025

Two new tools are now available in Vercel’s MCP server: get_access_to_vercel_url Generates a shareable URL that allows agent tools such as web fetch or Playwright to access deployments protected by Vercel Authentication. The URL is temporary and grants access without requiring login credentials. web_fetch_vercel_url Allows agents to directly fetch content from deployments protected by Vercel Authentication, even if a normal fetch would return 401 Unauthorized or 403 Forbidden. Get started with the Vercel MCP server. Read more

Node.js Vercel Functions now support fetch web handlers

August 19, 2025

Vercel Functions running on the Node.js runtime now support the fetch web handlers, improving interoperability across JavaScript runtimes and frameworks. You can still export individual HTTP methods, if preferred. Learn more about fetch web handlers in the docs. Read more

If agents are building your app, who gets the W-2?

August 18, 2025

Autonomous coding agents are not the future. They are already here. Agents can now design, build, test, and deploy an entire full-stack feature from front end to back end without a human touching the keyboard. The reality is that while this technology has advanced quickly, Generally Accepted Accounting Principles (GAAP) have not traditionally focused on the cost of tools used in development. Under current U.S. GAAP, you can capitalize certain third-party software costs if they are a direct cost of creating software during the application development stage. Historically, though, developer tools were treated as overhead because their cost could not be directly tied to capitalizable work. Under GAAP, work that meets the criteria should be capitalized. When agents perform that work, they should be treated no differently than salaried engineers. Read more

Native support for SvelteKit's new OpenTelemetry spans

August 18, 2025

Vercel now directly integrates with SvelteKit's new server-side OpenTelemetry spans. To get started, activate experimental tracing in SvelteKit: And create the tracing instrumentation file with the Vercel OpenTelemetry collector: Traces generated during tracing sessions will now include the built-in SvelteKit spans. You can also configure other collectors. See the SvelteKit observability docs for more information. Read more

Vercel Sandbox increases concurrency and port limits

August 18, 2025

Update (August 2026): These limits have since increased to 10,000 concurrent sandboxes for Pro and Enterprise teams, with dynamic vCPU allocation up to 5,000 per minute. See the announcement for details. Pro and Enterprise teams can now run up to 2,000 Vercel Sandboxes concurrently (up from 150), with each now able to expose up to 4 ports for external access. This enables larger traffic spikes for workloads like untrusted code execution, batch jobs, and automated testing, as well as more complex applications with multiple services or protocols running side-by-side. If you need a higher amount of concurrent sandboxes, you can contact our sales team to explore higher limits for your projects. Learn more in the Vercel Sandbox docs. Read more

Improved fake hardware detection with Vercel BotID

August 15, 2025

Vercel BotID Deep Analysis now uses an updated detection model that expands fingerprinting coverage for bespoke headless browsers and simulated device hardware. BotID is an invisible CAPTCHA that classifies sophisticated bots without interrupting real users. The new Deep Analysis model enables more accurate identification of stealthy automation frameworks and spoofed hardware profiles in real time. These updates take effect immediately for BotID Deep Analysis users with no action required, but we recommend upgrading to the latest botid@1.5.2. Get started with BotID today. Read more

The three types of AI bot traffic and how to handle them

August 13, 2025

AI bot traffic is growing across the web. We track this in real-time, and the data reveals three types of AI-driven crawlers that often work independently but together create a discovery flywheel that many teams disrupt without realizing it. Not all bots are harmful. Crawlers have powered search engines for decades, and we've spent just as long optimizing for them. Now, large language models (LLMs) need training data, and the AI tools built on them need timely, relevant updates. This is the next wave of discoverability and getting it right from the start can determine whether AI becomes a growth channel or a missed opportunity. Blocking AI crawlers today is like blocking search engines in the early days and then wondering why organic traffic vanishes. As users shift from Googling for web pages to prompting for direct answers and cited sources, the advantage will go to sites that understand each type of bot and choose where access creates value. Read more

The real serverless compute to database connection problem, solved

August 13, 2025

There is a long-standing myth that serverless compute inherently requires more connections to traditional databases. The real issue is not the number of connections needed during normal operation, but that some serverless platforms can leak connections when functions are suspended. In this post, we show why this belief is incorrect, explain the actual cause of the problem, and provide a straightforward, simple-to-use solution. Read more

How Coxwave delivers GenAI value faster with Vercel

August 13, 2025

Coxwave helps enterprises build GenAI products that work at scale. With their consulting arm, AX, and their analytics platform, Align, they support some of the world’s most technically sophisticated companies, including Anthropic, Meta, Microsoft, and PwC. Since the company’s founding in 2021, speed has been a defining trait. But speed doesn’t just mean fast models. For Coxwave, it means fast iteration, fast validation, and fast value delivery. To meet that bar, Coxwave reimagined their web app strategy with Next.js and Vercel. Read more

Introducing the Runtime Cache API

August 13, 2025

You can now access Vercel's Runtime Cache via API. The Runtime Cache is an ephemeral cache for storing and retrieving data across Functions, Routing Middleware, and Builds within the same region. It supports tag-based invalidation for precise and efficient cache control. You can get started with the API like this: You can monitor hit rates, invalidation patterns, and storage usage across your applications in the Observability dashboard's Runtime Cache tab. Runtime Cache reads and writes are billed regionally based on the runtime region. Learn more about Runtime Cache in the docs. Read more

Auto-recharge available in AI Gateway

August 12, 2025

Vercel AI Gateway now supports automatic credit recharging (top-ups), optionally refilling your balance before it runs out to keep your apps running without interruption. Auto-recharge is off by default and can be enabled or updated anytime in the AI Gateway dashboard or team billing settings. Set your top-up amount and trigger balance, optionally add a monthly spend limit, and your credits will automatically refill. Learn more about AI Gateway. Read more

Vercel's bot verification now supports Web Bot Auth

August 12, 2025

We collaborated with industry partners to advance the IETF proposal for Web Bot Auth and Vercel's bot verification system supports the new protocol. Now, Bot Protection can use HTTP Message Signatures to verify traffic automation from dynamic and distributed sources. Vercel maintains a comprehensive and actively curated directory of known bots that are verified by IP, reverse DNS, and now Web Bot Auth, which verifies bots via public-key cryptography in signed headers. This ensures that legitimate automation, like SEO crawlers, performance monitoring tools, and platform-integrated AI bots, can reliably access your site, while spoofed bots are blocked. Web Bot Auth's asymmetric signature proves the authenticity of the traffic regardless of its network origin, making it ideal for bots running in dynamic or serverless environments. Verified Bots using Web Bot Auth include signed headers to authenticate each request, allowing them to be recognized and allowed through Bot Protection and Challenge Mode. For example, ChatGPT operator signs its requests using Web Bot Auth, so is now allowed. Learn more about Bot Management. Read more

Vercel BotID now leverages Vercel's verified bot directory

August 12, 2025

Starting in botid@1.5.0, BotID’s Deep Analysis mode provides authenticated information for verified bots based on Vercel's directory of known and verified bots. This allows developers to detect verified bots in real time and make programmatic decisions based on bot identity. This allows you to securely allow known bots that are good for your business (such as agentic bots that purchase on behalf of users) while blocking other bots and sophisticated abuse. BotID is an invisible CAPTCHA that classifies sophisticated bots without interrupting real users. With this update, developers using Deep Analysis now get additional context about the bot itself, such as source IP range, reverse DNS, and user-agent validation, helping teams fine-tune how bots are handled before taking action. Get started with BotID and check out the documentation for verified bots in BotID. Read more

Cutting delivery times in half with v0

August 12, 2025

Ready.net is a core platform that helps utility companies manage their financing and compliance, the company works with a wide network of state-level stakeholders. New feature requirements come in fast, often vague, and always critical. With limited design resources supporting three teams, the company needed a way to speed up the loop between ideation, validation, and delivery. That’s where v0 came in. Read more

Claude Sonnet 4 now supports 1M token context in Vercel AI Gateway

August 12, 2025

You can now leverage Claude Sonnet 4's updated 1 million-token context window with Vercel's AI Gateway with no other provider accounts required. This release from Anthropic enables significantly larger inputs such as full codebases (~75,000+ lines) or large document sets. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to anthropic/claude-4-sonnet: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance and reliability to Claude Sonnet 4, AI Gateway leverages multiple model providers under the hood, including Anthropic and Bedrock. Learn more about AI Gateway and view the new AI Gateway model leaderboard. Read more

v0.dev -> v0.app

August 11, 2025

With a single prompt, anyone can go from idea to deployed app with UI, content, backend, and logic included. v0 is now agentic, helping you research, reason, debug, and plan. It can collaborate with you or take on the work end-to-end. From product managers writing specs to recruiters launching job boards, v0 is changing how teams operate. Read more

Add context when using Instant Rollback

August 11, 2025

You can now include a reason when performing an Instant Rollback. This message is visible to your team in the project overview and can include links or notes explaining the rollback. You can also update it at any time. Learn more about Instant Rollback. Read more

Cursor now supported on Vercel MCP

August 9, 2025

You can now use Cursor with Vercel MCP, our official Model Context Protocol (MCP) server. To ensure secure access, Vercel MCP currently supports AI clients that have been reviewed and approved by Vercel. With Vercel MCP you can explore projects, inspect failed deployments, fetch logs, and more, now all without leaving Cursor. To connect, either use click here for a one-click setup or add the following to your .cursor/mcp.json: Once added, Cursor will prompt you to log in with your Vercel account. Read more about using Cursor in Vercel MCP. Read more

How Zapier scales product partnerships with v0

August 8, 2025

Zapier is the leading AI orchestration platform, helping businesses turn intelligent insights into automated actions across nearly 8,000 apps. As AI tools and agents become more capable, Zapier provides the connective tissue to operationalize them, bridging the gap between decision and execution. Powered by Zapier extends this capability to partners. It enables SaaS and AI companies to embed Zapier’s automation engine directly into their products without needing to build or maintain thousands of integrations in-house. But explaining to partners what that experience can look like in their product was a challenge. Needing to move quickly with finite resources, the Zapier team could take a few weeks to design and build a clickable prototype. Now, with v0, the Powered by Zapier team can generate high-fidelity demos in just a few hours. The result: better conversations with partners, faster implementation cycles, and more integrations shipped for end users. Read more

vlt is now available in builds via zero configuration

August 8, 2025

Vercel now supports the vlt package manager for builds with zero configuration in builds. Starting today, Projects that contain a vlt-lock.json file will automatically run vlt install as the default Install Command using vlt. vlt requires node 20.x to run and is only available in the modern build image. Learn more about package manager support on Vercel. Read more

Bulk upgrade deprecated Node.js versions

August 8, 2025

Team owners and members can now upgrade all projects using Node.js 18 or earlier to Node.js 22 with one click in the Vercel Dashboard. This updates the Node.js version in project settings. If your project also defines a version in package.json, you'll need to update it manually. Existing deployments are not affected. View and upgrade deprecated Node.js projects now. Read more

Improved metrics search in Observability Plus

August 8, 2025

We’ve improved the metrics search and navigation experience in Vercel Observability, making it faster and easier to build custom queries. You can now: Quickly find metrics by typing partial names or common abbreviations like TTFB for "time to first byte" Browse all available metrics for an event in a side-by-side view Use keyboard shortcuts for faster navigation Access an optimized interface on mobile devices These updates are available now for all teams with Observability Plus. Try it out or learn more about Observability and Observability Plus. Read more

Vercel collaborates with OpenAI for GPT-5 launch

August 7, 2025

The GPT-5 family of models, released today, is now available through AI Gateway and in production on v0.dev. Thanks to OpenAI, Vercel has been testing these models over the past few weeks in v0, Next.js, AI SDK, and Vercel Sandbox. From our testing, GPT-5 is noticeably better at frontend design than previous models. It generates polished, balanced UIs with clean, composable code. Internally, we’ve already started using GPT-5 for Vercel's in-dashboard Agent and for v0.dev/gpt-5. GPT-5 shows strong performance in agent-based workflows. We have been impressed with it's long-context reasoning and ability to handle multiple tools in parallel have been especially effective in powering Vercel Agent. Read more

Vercel is the only vendor to be recognized as a Visionary in the 2025 Gartner® Magic Quadrant™ for Cloud-Native Application Platforms

August 7, 2025

At Vercel, we're building the platform that delivers every pixel and token, and powers every frontend, backend, and agent on the web. With more than 4M weekly active domains and 115B weekly requests served by Vercel, the most forward-thinking teams are choosing Vercel’s AI Cloud to deliver fast, secure, full-stack applications with zero friction, infinite scale, and complete developer freedom. We’re proud to be the only vendor named a Visionary in the 2025 Gartner® Magic Quadrant™ for Cloud Native Application Platforms. We believe that this recognition serves as validation: the future of the web is being built on Vercel. Read more

GPT-5, GPT-5-mini, and GPT-5-nano are now available in Vercel AI Gateway

August 7, 2025

You can now access GPT-5, GPT-5-mini, and GPT-5-nano by OpenAI, models designed to push the frontier of reasoning and domain expertise, using Vercel's AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to either openai/gpt-5 or openai/gpt-5-mini or openai/gpt-5-nano: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway and view the new AI Gateway model leaderboard. Read more

Introducing Vercel MCP: Connect Vercel to your AI tools

August 6, 2025

Today, we're launching the official Vercel MCP server, now in Public Beta. Vercel MCP is a secure, OAuth-compliant interface that lets AI clients interact with your Vercel projects. AI tools are becoming a core part of the developer workflow, but they've lacked secure, structured access to infrastructure like Vercel. With Vercel MCP, supported tools like Cursor and Claude can securely access logs, docs, and project metadata directly from within your development environment or AI assistant. Read more

Introducing AI Elements: Prebuilt, composable AI SDK components

August 6, 2025

AI Elements is a new open source library of customizable React components for building interfaces with the Vercel AI SDK. Built on shadcn/ui, it provides full control over UI primitives like message threads, input boxes, reasoning panels, and response actions. For example, you can use useChat from the AI SDK to manage state and streaming, and render responses using AI Elements. Getting started To install the components, you can initialize with our CLI, and pick your components, import them, and start building. Read the docs and start building better AI interfaces, faster. Read more

Microfrontends support is now in Public Beta

August 6, 2025

Microfrontends support is now available in Public Beta. Microfrontends allow you to split large applications into smaller ones so that developers can move more quickly. This support lets teams and large apps build and test independently, while Vercel assembles and routes the app into a single experience. This reduces build times, supports parallel development, and enables gradual legacy migration. Developers can use the Vercel Toolbar to iterate and test their apps independently, while navigations between microfrontends benefit from prefetching and prerendering for fast transitions between the applications. To get started with microfrontends, clone one of our examples or follow the quickstart guide: In the Vercel dashboard, navigate to the Microfrontends tab in Settings Create a microfrontends group containing all of your microfrontend projects Add the @vercel/microfrontends package to each microfrontend application Add a microfrontends.json configuration file to the default app, test in Preview, and deploy to Production when ready Learn more about microfrontends in our docs, or contact Vercel or your account team directly for more information. Read more

Claude Opus 4.1 is now supported in Vercel AI Gateway

August 5, 2025

You can now access Claude Opus 4.1, a new model released today, using Vercel's AI Gateway with no other provider accounts required. This release from Anthropic improves agentic task execution, real-world coding, and reasoning over the previous Opus 4 model. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to anthropic/claude-4.1-opus: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance and reliability to Claude Opus 4.1, AI Gateway leverages multiple model providers under the hood, including Anthropic and Bedrock. Learn more about AI Gateway and view the new AI Gateway model leaderboard. Read more

gpt-oss-20b and gpt-oss-120b are now supported in Vercel AI Gateway

August 5, 2025

You can now access gpt-oss-20b and gpt-oss-120b by OpenAI, open-weight reasoning models designed to push the open model frontier, using Vercel's AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to either openai/gpt-oss-20b or openai/gpt-oss-120b: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance and reliability to gpt-oss, AI Gateway leverages multiple model providers under the hood, including Groq, Baseten, Cerebras, and Huggingface. Learn more about AI Gateway and view the new AI Gateway model leaderboard. Read more

Vercel MCP now in Public Beta

August 4, 2025

Vercel’s official MCP (Model Control Protocol) server is now live at mcp.vercel.com in Public Beta. This server providing a remote interface with OAuth-based authorization for AI tools to securely interact with your Vercel projects. The server integrates with AI assistants, such as Claude.ai, Claude Code and Claude for desktop, and tools like VS Code, to: Search and navigate Vercel documentation Manage projects and deployments Analyze deployment logs Vercel MCP fully implements the latest MCP Authorization and Streamable HTTP specifications for enhanced security and performance. This update enhances collaboration between AI-driven workflows and Vercel ecosystems. For more details, read our the documentation. Read more

New custom visualization in Vercel Observability

August 4, 2025

Observability Plus users can now choose between line charts, volume charts, tables, or a big number when visualizing data returned by queries. Both the queries and their visualization settings can be saved to shareable notebooks. This update replaces fixed presets with customizable controls and is available now at no extra cost for teams on Observability Plus. Try it out or learn more about Observability and Observability Plus. Read more

v0: vibe coding, securely

August 4, 2025

Vibe coding has changed how software gets built. Tools like v0 make it possible to turn ideas into working prototypes in seconds. Anthropic's CEO predicts 90% of code will be AI-generated in 3-6 months. Adoption is accelerating fast, and for many builders, we're already there. But here's the uncomfortable truth: The faster you build, the more risk you create Last week, a viral app leaked 72k selfies and government IDs. This wasn’t a hack or advanced malware. It was caused by default settings, misused variables, and the absence of guardrails. A misconfigured Firebase bucket that was mistakenly left public for anyone to access. The app was built quickly, shipped without security review, and went viral. Read more

A new wave of software, shipped on Vercel

August 1, 2025

Shipped on Vercel showcases real apps in production on Vercel, built by teams rethinking how the web works. Read more

Deploy Hono backends with zero configuration

August 1, 2025

Vercel now natively supports Hono, a fast, lightweight backend framework built on web standards, with zero-configuration. With the code above, use Vercel CLI to develop and deploy your Hono application: With this improved integration, Vercel's framework-defined infrastructure now recognizes and deeply understands Hono applications, ensuring they benefit from optimizations made from builds, deployments, and application delivery. Now, new Hono applications deployed to Vercel benefit from Fluid compute, with Active CPU pricing, automatic cold start optimizations, background processing, and much more. Deploy Hono on Vercel or visit Hono's Vercel documentation. Read more

Vercel Open Source Program: Summer cohort

July 31, 2025

In April, we launched the Vercel Open Source Program, a developer initiative that gives maintainers the resources, credits, and support they need to ship faster and scale confidently, starting with the spring 2025 cohort. We're now honored to announce the summer 2025 cohort. From AI-powered calendars to beautifully styled React Native components, open source builders continue to amaze us. Here are the 28 projects from the summer cohort. Read more

AI SDK 5

July 31, 2025

With over 2 million weekly downloads, the AI SDK is the leading open-source AI application toolkit for TypeScript and JavaScript. Its unified provider API allows you to use any language model and enables powerful integrations into leading web frameworks. Read more

Z.ai's GLM-4.5 and GLM-4.5 Air are now supported in Vercel AI Gateway

July 29, 2025

You can now access GLM-4.5 and GLM-4.5 Air, new flagship models from Z.ai designed to unify frontier reasoning, coding, and agentic capabilities, using Vercel's AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to either zai/glm-4.5 or zai/glm-4.5-air: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. Learn more about AI Gateway. Read more

Join the v0 Ambassador Program

July 29, 2025

Since launch, we’ve seen a growing wave of people building with v0 and sharing what they’ve created, from full-stack apps to UI experiments. Now, we’re going a step further by sponsoring builders innovating and showcasing what’s possible with v0. Today we’re launching the v0 Ambassador Program as a way to recognize and enable members of our community who create, share, and inspire. Apply to join the v0 Ambassador Program and help others discover the magic of what's possible with v0. Read more

Fluid: How we built serverless servers

July 28, 2025

A few months ago, we announced Fluid compute, an approach to serverless computing that uses resources more efficiently, minimizes cold starts, and significantly reduces costs. More recently at Vercel Ship 2025, we introduced Active CPU pricing for even more cost-effective compute on Vercel. Fluid compute with Active CPU pricing powers over 45 billion weekly requests, saving customers up to 95% and never charging CPU rates for idle time. Behind the scenes, it took over two years to build the required infrastructure to make this possible. Read more

Generate shareable snapshots of Observability charts

July 28, 2025

You can now quickly share snapshots of any chart in Vercel Observability, making it easier to collaborate during debugging and incident response. Hover over a chart and press ⌘+C or Ctrl+C to copy a URL that opens a snapshot of the chart in Vercel Observability. The snapshot includes the same time range, filters, and settings as when copied. The link includes a preview image of the chart that unfurls in tools like Slack and Teams. Share links are public to ease sharing, but unguessable and ignored by search robots. Try it out or learn more about Observability and Observability Plus. Read more

Model Context Protocol (MCP) explained: An FAQ

July 25, 2025

Model Context Protocol (MCP) is a new way to help standardize the way large language models (LLMs) access data and systems, extending what they can do beyond their training data. It standardizes how developers expose data sources, tools, and context to models and agents, enabling safe, predictable interactions and acting as a universal connector between AI and applications. Instead of building custom integrations for every AI platform, developers can create an MCP server once and use it everywhere. Read more

Vercel and Solara6 partner to build better ecommerce experiences

July 25, 2025

Vercel is partnering with Solara6, a digital agency known for building high-performing ecommerce experiences for customers like Kate Spade, Coach, and Mattress Firm. Their work emphasizes AI-powered efficiencies, fast iteration cycles, and user experience, while prioritizing measurable outcomes. Solara6 customers see improvements in their developer velocity, operational costs, page load times, conversion rates, and organic traffic. Read more

Qwen3-Coder is now supported in Vercel AI Gateway

July 25, 2025

You can now access Qwen3 Coder, a model from QwenLM, an Alibaba Cloud company, designed to handle complex, multi-step coding workflows, using Vercel's AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to alibaba/qwen3-coder: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance and reliability to Qwen3 Coder, AI Gateway leverages multiple model providers under the hood, including Cerebras, DeepInfra, and Parasail. Learn more about AI Gateway. Read more

Growthbook joins the Vercel Marketplace

July 24, 2025

GrowthBook, the open-source experimentation platform, is now available as a native integration on the Vercel Marketplace. Easily add feature flags and A/B testing to your Vercel projects with minimal setup. With GrowthBook on Vercel, you can: Declare flags in code using Flags SDK and the @flags-sdk/growthbook adopter Sync feature flags directly to Vercel Edge Config, powering low latency evaluation Bring your own data using GrowthBook’s warehouse-native A/B testing platform Explore the Template to view and deploy the example, with one-click setup and unified billing. Read more

Build your own AI app builder with the v0 Platform API

July 23, 2025

The v0 Platform API is a text-to-app API that gives developers direct access to the same infrastructure powering v0.dev. Currently in beta, the platform API exposes a composable interface for developers to automate building web apps, integrate code generation into existing features, and build new products on top of LLM-generated UIs. Read more

Vercel BotID now available for all frameworks

July 22, 2025

You can now use Vercel BotID to protect your most sensitive endpoints in any JavaScript framework, like SvelteKit and Nuxt. BotID is our advanced bot protection for high-value endpoints like registration, checkout, and AI interactions. Since launch, it has already protected nearly a million API requests. Installing or upgrading to botid@1.4.3 adds support for universal JavaScript environments with the new initBotId({ protect: ... }) function. Here's an example of initBotId used to set up BotID in SvelteKit: Check out the updated documentation for setup instructions across all supported frameworks. Read more

Transform rules are now available in vercel.json

July 22, 2025

You can now define transform rules in vercel.json to modify HTTP request and response headers or query parameters, without changing application code. Unlimited transform rules are available for all customers, and let you: Set, append, or delete request headers, response headers, and query parameters Use conditional logic to apply changes based on request metadata Match by equality, inequality, prefixes, suffixes, inclusion in string arrays, or numeric comparisons for fine-grained control This expands the flexibility of Vercel's CDN, which already supports routing behavior like redirects and rewrites to external origins. For example: Refer to the transform rules documentation for detailed examples. Read more

OpenAI-compatible API endpoints now supported in AI Gateway

July 21, 2025

You can now use OpenAI-compatible client libraries and tools with AI Gateway through a simple URL change, allowing you to access 100s of models with no code rewrites required. Here is a Python example with the OpenAI client library: This makes it easy to keep your current tools and workflows while improving uptime, tokens per minute, quotas, and reliability via provider failover and adding observability through the AI Gateway. Learn more in the AI Gateway docs and see more examples here. Read more

Open Vercel documentation pages in AI providers

July 18, 2025

You can now copy Vercel documentation pages as markdown, or open them directly in v0, Claude or ChatGPT. This allows you to use documentation content as context when working with AI tools. Visit any documentation page and use the dropdown in the top right of the page. Using the copy page dropdown Navigate to any documentation page Click the copy page dropdown in the top right corner Select your provider or copy as markdown The page content will be formatted and loaded into the selected AI provider. Read more

Grep a million GitHub repositories via MCP

July 17, 2025

Grep now supports the Model Context Protocol (MCP), enabling AI apps to query a million public GitHub repositories using a standard interface. Whether you're building in Cursor, using Claude, or integrating your own agent, Grep can now serve as a searchable code index over HTTP. Read more

Moonshot AI's Kimi K2 model is now supported in Vercel AI Gateway

July 15, 2025

You can now access Kimi K2, a new mixture-of-experts (MoE) language model from Moonshot AI, using Vercel's AI Gateway with no other provider accounts required. AI Gateway lets you call the model with a consistent unified API and just a single string update, track usage and cost, and configure performance optimizations, retries, and failover for higher than provider-average uptime. To use it with the AI SDK v5, start by installing the package: Then set the model to moonshotai/kimi-k2: Includes built-in observability, Bring Your Own Key support, and intelligent provider routing with automatic retries. To deliver high performance and reliability to Kimi K2, AI Gateway leverages multiple model providers under the hood, including direct to Moonshot AI, Groq, DeepInfra, Fireworks AI, and Parasail. Learn more about AI Gateway. Read more

OAuth support added to MCP Adapter

July 15, 2025

Secure your MCP servers with OAuth using version 1.0.0 of the MCP Adapter, which now includes official support for the MCP Authorization spec. This release introduces: Helper functions for OAuth-compliant authorization flows A new withMcpAuth wrapper for securing routes One-click deployable examples with popular auth providers like Better Auth, Clerk, Descope, Stytch, and WorkOS Here’s an example of how to integrate auth in your MCP server: Additionally, use the protectedResourceHandler to expose resource server metadata for compliant clients. Learn more in the MCP Auth documentation. Start building secure MCP servers Deploy an example MCP server by cloning our Next.js MCP template, or explore starter integrations from our auth partners: Better Auth Clerk Descope Stytch WorkOS Read more

Search any public GitHub repo with Grep

July 14, 2025

You can now use Grep to search any public repository on GitHub, no longer limited to the 1M+ pre-indexed repos. To search a specific repo, use grep.app/[owner]/[repo]. For example: visit grep.app/vercel/ai and start typing a search query (try streamText). Get quick, full-text and regular expression search across the repo without any setup. Read more

Clerk joins the Vercel Marketplace

July 14, 2025

Clerk is now available as an authentication provider on the Vercel Marketplace. Built for modern frameworks like Next.js, Clerk simplifies authentication while giving teams full control over UI, sessions, and user roles, all tightly integrated with Vercel’s deployment model. With the integration, you get access to: Instant provisioning of Clerk apps from your Vercel dashboard Complete user management with hosted dashboards, sessions, and roles Built-in and scalable billing and subscription management Get started with Clerk on the Vercel Marketplace. Read more

More Secure Deployment Protection

July 14, 2025

Deployment Protection safeguards preview and production URLs so that users can't access the domains that you don't want them to. Starting today, the Standard Deployment Protection option has been updated for new projects to protect all automatically generated domains, including the production branch git domain (for example project-git-main.vercel.app). Existing projects can update to this new behavior in the Project settings page in the Vercel dashboard. Read more

The AI Cloud: A unified platform for AI workloads

July 10, 2025

For over a decade, Vercel has helped teams develop, preview, and ship everything from static sites to full-stack apps. That mission shaped the Frontend Cloud, now relied on by millions of developers and powering some of the largest sites and apps in the world. Now, AI is changing what and how we build. Interfaces are becoming conversations and workflows are becoming autonomous. We've seen this firsthand while building v0 and working with AI teams like Browserbase and Decagon. The pattern is clear: developers need expanded tools, new infrastructure primitives, and even more protections for their intelligent, agent-powered applications. At Vercel Ship, we introduced the AI Cloud: a unified platform that lets teams build AI features and apps with the right tools to stay flexible, move fast, and be secure, all while focusing on their products, not infrastructure. Read more

Vercel Blob now available in all Vercel Regions

July 10, 2025

You can now create Vercel Blob stores in any of the 19 Vercel Regions. Selecting a region closer to your Functions and users allows you to optimize upload speed and comply with data-residency requirements when required. Selecting a region is available at creation time in the Vercel dashboard or when using the Vercel CLI (version 44.3.0). Learn more about Vercel Blob in the documentation. Read more

v0 Platform API now in beta

July 9, 2025

The v0 Platform API is now available in public beta. The v0 Platform API is a text-to-app API — it provides programmatic access to v0’s app generation pipeline: Generating code for web apps from prompts Structured parsing of generated code Automatic error fixing Link with a rendered preview This API also supports programmatic control of v0.dev, including creating and managing both chats and projects. We'll be bringing more of v0.dev's functionality into the Platform API soon. The v0 Platform API is designed for integration into development workflows, automation scripts, and third-party tools. Check out our TypeScript SDK and documentation to get started. Read more

Web Application Firewall control now available with vercel.json

July 9, 2025

You can now control Vercel’s Web Application Firewall (WAF) actions directly in vercel.json, alongside existing support in the dashboard, API, and terraform. This approach provides a structured way for both developers and agents to declaratively define and push rules to projects. Agents can use code-generating prompts to author new rules that are easily injected into the project’s vercel.json. The has and missing matchers have also been enhanced to support more expressive conditions across headers, rewrites, redirects, and routes. Matching options include: String equality and inequality Regular expressions Prefixes and suffixes Inclusion and exclusion from string arrays Numeric comparisons The following example shows how to deny a request that is prefixed by a specific header: Read more about Vercel's WAF and configuring WAF rules in vercel.json. Read more

Inngest joins the Vercel Marketplace

July 9, 2025

You can now install Inngest directly from the Vercel Marketplace to quickly build reliable background jobs and AI workflows for your Next.js app. Inngest is a great fit to add AI features or emerging agentic patterns to your Vercel projects: Write background jobs directly in your `app/` directory Full support for preview environments and branching One-click install and integrated billing with a generous free tier (100K executions/month) Start building workflows with Inngest on the Vercel Marketplace today. Read more

NuxtLabs joins Vercel

July 8, 2025

NuxtLabs, creators and stewards of Nitro and Nuxt, are joining Vercel. Read more

Sandbox now supports sudo and installing RPM packages

July 4, 2025

You can now run commands with sudo inside Vercel Sandbox, giving you full control over runtime environment setup, just like on a traditional Linux system. This makes it possible to install system dependencies at runtime, like Go, Python packages, or custom binaries, before executing your code. sudo is available via the runCommand method: The sandbox sudo configuration is designed to be easy to use: PATH is preserved HOME is set to /root Custom environment variables like env: { FOO: "bar" } are passed through With sudo on Sandbox it's easier to run untrusted code in isolated environments with the right permissions, with no workarounds required. Learn more about Vercel Sandbox and sudo in the documentation. Read more

Correlate logs and traces with OpenTelemetry in Vercel Log Drains

July 4, 2025

Vercel now automatically correlates logs with distributed traces for customers using OpenTelemetry to instrument their applications. Traces are a way to collect data about the performance and behavior of your application and help identify the cause of performance issues, errors, and other problems. OpenTelemetry (OTel) is an open source project that allows you to instrument your application to collect traces. When a request is traced using OTel, Vercel will enrich the relevant logs with trace and span identifiers. This allows you to correlate your individual logs to a trace or span. This feature is available to customers using log drains through our integrations with Datadog and Dash0. No action is required and log to trace correlation will happen automatically going forward for customers using OTel with any of these integrations. Learn more about correlating logs to traces using log drains. Read more

CVE-2025-49005

July 3, 2025

Summary A cache poisoning vulnerability affecting Next.js App Router >=15.3.0 < 15.3.3 / Vercel CLI 41.4.1–42.2.0 has been resolved. The issue allowed page requests for HTML content to return a React Server Component (RSC) payload instead under certain conditions. When deployed to Vercel, this would only impact the browser cache, and would not lead to the CDN being poisoned. When self-hosted and deployed externally, this could lead to cache poisoning if the CDN does not properly distinguish between RSC / HTML in the cache keys. Impact Under specific conditions involving App Router, middleware redirects, and omitted Vary headers, applications may: Serve RSC payloads in place of HTML Cache these responses at the browser or CDN layer Display broken or incorrect client content This issue occurs in environments where middleware rewrites or redirects result in improper cache key separation, because the cache-busting parameter added by the framework is stripped by the user’s redirect. Resolution The issue was resolved in Next.js 15.3.3 by: Ensuring the Vary header is correctly set to distinguish between different content types Customers hosting on Vercel with deployments that used the imp

CVE-2025-49826

July 3, 2025

Summary A vulnerability affecting Next.js has been addressed. It impacted versions >=15.1.0 <15.1.8 and involved a cache poisoning bug leading to a Denial of Service (DoS) condition. Impact This issue does not impact customers hosted on Vercel. Under certain conditions, this issue may allow a HTTP 204 response to be cached for static pages, leading to the 204 response being served to all users attempting to access the page. This issue required the below conditions to be exploitable: Using an affected version of Next.js, and; A route using cache revalidation with ISR (next start or standalone mode); and A route using SSR, with a CDN configured to cache 204 responses. Resolution The issue was resolved by removing the problematic code path that would have caused the 204 response to be set. Additionally, we removed the race condition that could have led to this cache poisoning by no longer relying on a shared response object to populate the Next.js response cache. Credit Thanks to Allam Rachid (zhero) and Allam Yasser (inzo_) for responsible disclosure. References GHSA-67rr-84xm-4c7r Read more

New usage dashboard for Pro customers

July 3, 2025

Pro teams can now access a new usage dashboard (recently introduced to Enterprise customers) with improved filtering, detailed breakdowns, and export options to better understand usage and costs by product and project. You can now break down usage by: Product to quickly identify usage, drill down into spikes, and track costs of a single or set of products Team and project to understand your costs and monitor team activity across all or specific apps CSV exports for external analysis via integration into your cost observability tools and spreadsheets Explore the new dashboard today. Read more

Zero-configuration support for Nitro

July 3, 2025

Vercel now supports Nitro applications, a backend toolkit for building web servers, with zero-configuration. Nitro powers frameworks like Nuxt.js, TanStack Start, and SolidStart. Deploy Nitro on Vercel or visit Nitro's Vercel documentation. Read more

New webhook events for domain management

June 26, 2025

You can now subscribe to webhook events for deeper visibility into domain operations on Vercel. New event categories include: Domain transfers: Track key stages in inbound domain transfers. Domain renewals: Monitor renewal attempts and auto-renew status changes, ideal for catching failures before they impact availability. Domain certificates: Get notified when certificates are issued, renewed, or removed, helping you maintain valid HTTPS coverage across environments. DNS changes: Receive alerts when DNS records are created, updated, or deleted. Project Domain Management: Detect domain lifecycle changes across projects, including creation, updates, verification status, and reassignment. These events are especially valuable for multi-tenant platforms that dynamically assign domains per user or customer. They also help teams build monitoring and alerting into critical domain and certificate operations. For details on how to subscribe, visit the webhook documentation. Read more

Vercel Ship 2025 recap

June 26, 2025

My first week at Vercel coincided with something extraordinary: Vercel Ship 2025. Vercel Ship 2025 showcased better building blocks for the future of app development. AI has made this more important than ever. Over 1,200 people gathered in NYC for our third annual event, to hear the latest updates in AI, compute, security, and more. Read more

Rolling Releases are now generally available

June 25, 2025

Rolling Releases are now generally available, allowing safe, incremental rollouts of new deployments with built-in monitoring, rollout controls, and no custom routing required. Each rollout starts at a defined stage and can either progress automatically or be manually promoted to a full release. You can configure rollout stages per project and decide how each stage progresses, with updates propagating globally in under 300ms through our fast propagation pipeline. Rolling releases also include: Real-time monitoring: Track and compare error rates and Speed Insights (like Core Web Vitals, Time to First Byte, and more) between versions Flexible controls: Rollouts can be managed via REST API, CLI, the project dashboard, or the Vercel Terraform provider Version-labeled logs: Logs and telemetry are labeled by deployment for easier debugging Pro and Enterprise teams can enable Rolling Releases on one project at no additional cost. Enterprise customers can upgrade to unlimited projects. Learn more about Rolling Releases or enable it on your project. Read more

Higher defaults and limits for Vercel Functions running Fluid compute

June 25, 2025

The default limits for Vercel Functions using Fluid compute have increased, with longer execution times, more memory, and more CPU. The default execution time, for all projects on all plans, is now 300 seconds (5 minutes): Plan Default Maximum Hobby 300s (previously 60s) 300s (previously 60s) Pro 300s (previously 90s) 800s Enterprise 300s (previously 90s) 800s Pro and Enterprise teams using supported Node.js and Python runtimes can now opt into an extended maximum of 1800s (30 minutes, beta). See the June 15, 2026 update. Memory and CPU instance sizes have also been updated: Standard (default) is now 1 vCPU / 2 GB (previously 1 vCPU / 1.7 GB) Performance is now 2 vCPU / 4 GB (previously 1.7 vCPU / 3 GB) These increased instances are enabled by Active CPU pricing, which charges based on actual compute time. Periods of memory-only usage are billed at a significantly lower rate, making longer executions more cost-efficient. You can view logs to determine if your functions are hitting execution limits and adjust the max duration or upgrade your plan as needed. Learn more about Vercel Function limits. Read more

Edge Middleware and Edge Functions are now powered by Vercel Functions

June 25, 2025

Note: This changelog entry is historical. Edge Functions are deprecated for new projects. Use Vercel Functions with the Node.js runtime and Fluid compute. Use Routing Middleware for request-time routing before a response completes. Functions using the Edge runtime now run on the unified Vercel Functions infrastructure. This applies to both before and after the cache: Routing Middleware is now Vercel Routing Middleware, a new infrastructure primitive that runs full Vercel Functions with Fluid compute before the cache Edge Functions are now Vercel Functions using the Edge Runtime after the cache With these changes, all functions including those running the Edge runtime are: Fluid compute-ready: Runs on Fluid compute for better performance and cost efficiency Multi-runtime: Supports Node.js and Edge runtimes Framework-driven: Deployed automatically from supported framework code Consistent pricing: Uses unified Vercel Functions pricing based on Active CPU time across all compute types Vercel Routing Middleware is now generally available to all users. Learn more about Routing Middleware. Read more

Run untrusted code with Vercel Sandbox

June 25, 2025

Vercel Sandbox is a secure cloud resource powered by Fluid compute. It is designed to run untrusted code, such as code generated by AI agents, in isolated and ephemeral environments. Sandbox is a standalone SDK that can be executed from any environment, including non-Vercel platforms. Sandbox workloads run in ephemeral, isolated microVMs via the new Sandbox SDK, supporting execution times up to 45 minutes. Sandbox uses the Fluid compute model and charges based on Fluid’s new Active CPU time, meaning you only pay for compute when actively using CPU. See Sandbox pricing for included allotments and pricing for Hobby and Pro teams. Now in Beta and available to customers on all plans. Learn more about Vercel Sandbox. Read more

Lower pricing with Active CPU pricing for Fluid compute

June 25, 2025

Vercel Functions on Fluid Compute now use Active CPU pricing, which charges for CPU only while it is actively doing work. This eliminates costs during idle time and reduces spend for workloads like LLM inference, long-running AI agents, or any task with idle time. Active CPU pricing is built on three core metrics: Active CPU: Time your code is actively executing in an instance. Priced at $0.128 per hour Provisioned Memory: Memory allocated to the instance, billed at a lower rate. Priced at $0.0106 per GB-Hour Invocations: One charge per function call An example of this in action: A function running Standard machine size at 100% active CPU would now cost ~$0.149 per hour (1 Active CPU hour + 2 GB of provisioned memory). Previously this would have cost $0.31842 per hour (1.7 GB Memory × $0.18). Active CPU pricing is now enabled by default for all Hobby, Pro, and new Enterprise teams. For existing Enterprise customers, availability depends on your current plan configuration. This change takes effect after a redeploy. Learn more about Fluid compute with Active CPU pricing and read the announcement. Read more

AI Gateway is now in Beta

June 25, 2025

AI Gateway gives you a single endpoint to access a wide range of AI models across providers, with better uptime, faster responses, no lock-in. Now in Beta, developers can use models from providers like OpenAI, xAI, Anthropic, Google, and more with: Usage-based billing at provider list prices Bring-Your-Own-Key support Improved observability, including per-model usage, latency, and error metrics Simplified authentication Fallback and provider routing for more reliable inference Higher throughput and rate limits Try AI Gateway for free or check out the documentation to learn more. Read more

​Introducing BotID, invisible bot filtering for critical routes

June 25, 2025

Modern sophisticated bots don’t look like bots. They execute JavaScript, solve CAPTCHAs, and navigate interfaces like real users. Tools like Playwright and Puppeteer can script human-like behavior from page load to form submission. Traditional defenses like checking headers or rate limits aren't enough. Bots that blend in by design are hard to detect and expensive to ignore. Enter BotID: A new layer of protection on Vercel. Think of it as an invisible CAPTCHA to stop browser automation before it reaches your backend. It’s built to protect critical routes where automated abuse has real cost such as checkouts, logins, signups, APIs, or actions that trigger expensive backend operations like LLM-powered endpoints. Read more

Introducing Active CPU pricing for Fluid compute

June 25, 2025

Fluid compute exists for a new class of workloads. I/O bound backends like AI inference, agents, MCP servers, and anything that needs to scale instantly, but often remains idle between operations. These workloads do not follow traditional, quick request-response patterns. They’re long-running, unpredictable, and use cloud resources in new ways. Fluid quickly became the default compute model on Vercel, helping teams cut costs by up to 85% through optimizations like in-function concurrency. Today, we’re taking the efficiency and cost savings further with a new pricing model: you pay CPU rates only when your code is actively using CPU. Read more

Vercel Queues is now in Limited Beta

June 25, 2025

Vercel Queues is a message queue service built for Vercel applications, in Limited Beta. Vercel Queues lets you offload work by sending tasks to a queue, where they’ll be processed in the background. This means users don’t have to wait for slow operations to finish during a request, and your app can handle retries and failures more reliably. Under the hood, Vercel Queues uses an append-only log to store messages and ensures tasks such as AI video processing, sending emails, or updating external services are persisted and never lost. Key features of Vercel Queues: Pub/Sub pattern: Topic-based messaging allowing for multiple consumer groups Streaming support: Handle payloads without loading them entirely into memory Streamlined auth: Automatic authentication via OIDC tokens SDK: TypeScript SDK with full type safety If you have any questions, let us know in the Vercel Community. Read more

Vercel Agent now in Limited Beta

June 25, 2025

Vercel Agent is now available in Limited Beta. Agent is an AI assistant built into the Vercel dashboard that analyzes your app performance and security data. Agent focuses on Observability, summarizing anomalies, identifying likely causes, and recommending specific actions. These actions can span across the platform, including managing firewall rules in response to traffic spikes or geographic anomalies, and identifying optimization opportunities within your application. Insights appear contextually as detailed notebooks with no configuration required. Sign up with Vercel Community, express your interest in participating, and we'll reach out to you. Read more

Vercel BotID is now generally available

June 25, 2025

Vercel BotID is an invisible CAPTCHA with no visible challenges or manual bot management required. BotID is a new protection layer on Vercel designed for public, high-value routes such as checkouts, signups, AI chat interfaces, LLM-powered endpoints, and public APIs that are targets for sophisticated bots mimicking real user behavior. Unlike IP-based or heuristic systems, BotID: Silently collects thousands of signals that distinguish human users from bot Mutates these detections on every page load, evading reverse engineering and sophisticated bypasses Streams attack data into a global machine learning mesh, collectively strengthening protection for all customers Powered by Kasada, BotID integrates into your application with a type-safe SDK: Client-side detection using the component Server-side verification with the checkBotId function Automatic labeling of logs and telemetry for blocked sessions BotID traffic is visible in the Firewall dashboard and can be filtered by verdict (pass or fail), user agent, country, IP address, request path, target path, JA4 digest, and host. Read the announcement or documentation to learn more, or try BotID today. Read more

Vercel Microfrontends is now in Limited Beta

June 25, 2025

Vercel Microfrontends is now available in Limited Beta for Enterprise teams, enabling you to deploy and manage multiple frontend applications that appear as one cohesive application to users. This allows you to split large applications into smaller, independently deployable units that each team can build, test, and deploy using their own tech stack, while Vercel handles integration and routing across the platform. Faster development for large apps: Smaller units reduce build times and enable teams to move independently Independent team workflows: Each team manages its own deployment pipeline and framework Incremental migration: Modernize legacy systems piece by piece without slow, large-scale rewrites Learn more about Vercel Microfrontends. Reach out to your account representative or contact sales to join the limited beta. Read more

Manually purge the CDN cache

June 24, 2025

Users with the Member role can now purge Vercel’s CDN cache manually, either via the project's cache settings dashboard or by running vercel cache purge --type=cdn in CLI version 44.2.0 or later. By default, the CDN cache is purged automatically with each new deployment. For cases where you want to refresh cached content instantly (without waiting for a new build), you can now manually purge the global CDN cache in milliseconds. This is especially useful for persistent cache scenarios, like Image Optimization, where paths are cached across deployments. If upstream images have changed, you can now force a refresh instantly. Learn more in the documentation. Read more

Vercel Blob CLI is now available

June 24, 2025

The Vercel CLI (version 43.3.0) now includes Blob commands, allowing you to manage your Vercel Blob stores and files directly from the terminal. Learn more about the Vercel Blob CLI and Vercel Blob. Read more

WPP and Vercel: Bringing AI to the creative process

June 24, 2025

Today, we're announcing an expansion of our partnership with WPP. A first-of-its-kind agency collaboration that now brings v0 and AI SDK directly to WPP's global network of creative teams and their clients. Read more

Keith Messick joins Vercel as CMO

June 23, 2025

Vercel is evolving to meet the expanding potential of AI while staying grounded in the principles that brought us here. We're extending from frontend to full stack, deepening our enterprise capabilities, and powering the next generation of AI applications, including integrating AI into our own developer tools. Today, we’re welcoming Keith Messick as our first Chief Marketing Officer to support this growth and (as always) amplify the voice of the developer. Read more

Find teams, projects, and pages in the Vercel dashboard with universal search

June 23, 2025

There is now a search feature in the top right corner of every page in the vercel.com dashboard. This search allows you to instantly find: Teams Projects Deployments (by branch) Pages Settings For more complex queries you can also ask the Navigation Assistant. This AI-powered feature can locate any page in the dashboard and apply filters based on your question. Learn more about Find in the documentation. Read more

Turso Cloud joins the Vercel Marketplace

June 20, 2025

Turso now offers a native integration with Vercel, available as Database & Storage provider in the Marketplace. The Turso integration brings fast, distributed SQLite databases to your Vercel projects with: Seamless integration with Vercel, including one-click setup and unified billing Create unlimited SQLite databases in the cloud for production workloads. Serverless access or sync. A developer-friendly experience, configurable through Vercel CLI workflows Get started with Turso on the Vercel Marketplace. Read more

Two-factor authentication (2FA) team enforcement

June 19, 2025

Teams can now require all members to enable two-factor authentication (2FA) for added security. Team owners can enable enforcement in the Security & Privacy section of team settings. Owner controls View and filter each member’s 2FA status in the team members settings Member restrictions Once enforcement is enabled, members without 2FA will be restricted from: Triggering builds from pull requests Accessing new preview deployments Viewing the team dashboard Making API requests Using access tokens Enforcement lock-in & visibility Members of a team with 2FA enforcement cannot disable 2FA unless they leave the team In each user’s account settings, teams that require 2FA are now listed for clarity Enable 2FA enforcement today, and learn more in our docs. Read more

Create and share queries with notebooks in Vercel Observability

June 19, 2025

Observability Plus users can now create a collection of queries in notebooks to collaboratively explore their observability data. Queries in Vercel Observability allow you to explore log data and visualize traffic, performance, and other key metrics, and can now be saved to notebooks. By default, notebooks are only visible to the user who created the notebook, but you have the option to share a notebook with all members of your team. This is available to Observability Plus subscribers at no additional cost. Try it out or learn more about Observability and Observability Plus. Read more

Tray.ai cut build times from a day to minutes with Vercel

June 16, 2025

Tray.ai is a composable AI integration and automation platform that enterprises use to build smart, secure AI agents at scale. To modernize their marketing site, they partnered with Roboto Studio to migrate off their legacy solution and outdated version of Next.js. The goal: simplify the architecture, consolidate siloed repos, and bring content and form management into one unified system. After moving to Vercel, builds went from a full day to just two minutes. Read more

Introducing the Dubai Vercel region (dxb1)

June 16, 2025

Dubai (dxb1) is now part of Vercel’s delivery network, extending our global CDN's caching and compute to reduce latency for users in the Middle East, Africa, and Central Asia without requiring any changes. The new Dubai region serves as the first stop for end-users based on proximity and network conditions. It's generally available and serving billions of requests. Teams can configure Dubai as an execution region for Vercel Functions, which supports Fluid compute to increase resource and cost efficiency, minimize cold starts, and scale dynamically with demand. Learn more about Vercel Regions and Dubai's regional pricing. Read more

Building efficient MCP servers

June 12, 2025

The Model Context Protocol (MCP) standardizes how to build integrations for AI models. We built the MCP adapter to help developers create their own MCP servers using popular frameworks such as Next.js, Nuxt, and SvelteKit. Production apps like Zapier, Composio, Vapi, and Solana use the MCP adapter to deploy their own MCP servers on Vercel, and they've seen substantial growth in the past month. MCP has been adopted by popular clients like Cursor, Claude, and Windsurf. These now support connecting to MCP servers and calling tools. Companies create their own MCP servers to make their tools available in the ecosystem. The growing adoption of MCP shows its importance, but scaling MCP servers reveals limitations in the original design. Let's look at how the MCP specification has evolved, and how the MCP adapter can help. Read more

Improved team overview page

June 12, 2025

We've improved the team overview in the Vercel dashboard: Activity is now sorted by your activity only Projects can be filtered by git repository Usage for the team is now shown as a card on the overview directly To learn more about the Vercel dashboard, visit the documentation. Read more

Improved unhandled Node.js errors in Fluid compute

June 12, 2025

Fluid compute now gracefully handles uncaught exceptions and unhandled rejections in Node.js by logging the error, allowing inflight requests to complete, and then exiting the process. This prevents concurrent requests running on same fluid instance from being inadvertently terminated in case of unhandled errors, providing the isolation of traditional serverless invocations. Enable Fluid for your existing projects, and learn more in our blog and documentation. Read more

Designing and building the Vercel Ship conference platform

June 11, 2025

Our two conferences (Vercel Ship and Next.js Conf) are our chance to show what we've been building, how we're thinking, and cast a vision of where we're going next. It's also a chance to push ourselves to create an experience that builds excitement and reflects the quality we strive for in our products. For Vercel Ship 2025, we wanted that experience to feel fluid and fast. This is a look at how we made the conference platform and visuals, from ferrofluid-inspired 3D visuals and generative AI workflows, to modular component systems and more. Read more

Filter runtime logs for fatal function errors

June 10, 2025

You can now filter runtime logs to view fatal function errors, such as Node.js crashes, using the Fatal option in the levels filter. When a log entry corresponds to a fatal error, the right-hand panel will display Invocation Failed in the invocation details. Try it out or learn more about runtime logs Read more

How we’re adapting SEO for LLMs and AI search

June 10, 2025

Search is changing. Backlinks and keywords aren’t enough anymore. AI-first interfaces like ChatGPT and Google’s AI Overviews now answer questions before users ever click a link (if at all). Large language models (LLMs) have become a new layer in the discovery process, reshaping how, where, and when content is seen. This shift is changing how visibility works. It’s still early, and nobody has all the answers. But one pattern we're noticing is that LLMs tend to favor content that explains things clearly, deeply, and with structure. "LLM SEO" isn’t a replacement for traditional search engine optimization (SEO). It’s an adaptation. For marketers, content strategists, and product teams, this shift brings both risk and opportunity. How do you show up when AI controls the first impression, but not lose sight of traditional ranking strategies? Here’s what we’ve noticed, what we’re trying, and how we’re adapting. Read more

Building secure AI agents

June 9, 2025

An AI agent is a language model with a system prompt and a set of tools. Tools extend the model's capabilities by adding access to APIs, file systems, and external services. But they also create new paths for things to go wrong. The most critical security risk is prompt injection. Similar to SQL injection, it allows attackers to slip commands into what looks like normal input. The difference is that with LLMs, there is no standard way to isolate or escape input. Anything the model sees, including user input, search results, or retrieved documents, can override the system prompt or event trigger tool calls. If you are building an agent, you must design for worst case scenarios. The model will see everything an attacker can control. And it might do exactly what they want. Read more

v0-1.5-md & v0-1.5-lg now in beta on the Models API

June 9, 2025

Beta access is now available for v0-1.5-md (128K token context) and v0-1.5-lg (512K token context) on our Models API. For full details and examples, see the Models API docs: https://vercel.com/docs/v0/api Read more

Observability added to AI Gateway alpha

June 9, 2025

The AI Gateway, currently in alpha for all users, lets you switch between ~100 AI models without needing to manage API keys, rate limits, or provider accounts. Vercel Observability now includes a dedicated AI section to surface metrics related to the AI Gateway. This update introduces visibility into: Requests by model Time to first token (TTFT) Request duration Input/output token count Cost per request (free while in alpha) You can view these metrics across all projects or drill into per-project and per-model usage to understand which models are performing well, how they compare on latency, and what each request would cost in production. Learn more about Observability. Read more

Claude Code and Cursor Agent no longer require a team seat

June 6, 2025

We've updated our build logic to ensure Git commits authored by Claude Code or Cursor Agent can trigger deployments on Vercel. A team seat is not required. If your agent encounters any issues building on Vercel, please contact us. Read more

Bot Protection is now generally available

June 5, 2025

Vercel Web Application Firewall's Bot Protection managed ruleset is now generally available for all users, at no additional cost. Bot Protection helps reduce automated traffic from non-browser sources and allows you to respond based on two action choices: Log Only Action: Logs identified bot traffic in the Firewall tab without blocking requests Challenge Action: Serves a browser challenge to traffic from non-browser sources. Verified bots are automatically excluded During the beta period, Bot Protection challenged over 650 million requests of potential non-browser requests. Bot Protection complements Vercel's existing mitigations, which already block common threats like DDoS attacks, low quality traffic, and spoofed traffic. It adds an extra layer of protection for any automated traffic that is not clearly malicious. Learn more about the Bot Protection managed ruleset and the Vercel Firewall. If you'd like your bot to be verified as well, head over to bots.fyi. Read more

Pre-generate SSL certs, now in the Domains dashboard

June 5, 2025

You can now pre-generate SSL certificates directly from the Vercel Domains dashboard, enabling zero-downtime domain migrations without using the CLI. After adding an existing domain to your project, select Pre-Generate Certificate to issue certificates before updating DNS records and initiating the remainder of your domain migration. You can still import a zone file or use Domain Connect to migrate DNS records from your previous provider. Try it out or learn more in the docs. Read more

The no-nonsense approach to AI agent development

June 4, 2025

AI agents are software systems that take over tasks made up of manual, multi-step processes. These often require context, judgment, and adaptation, making them difficult to automate with simple rule-based code. While traditional automation is possible, it usually means hardcoding endless edge cases. Agents offer a more flexible approach. They use context to decide what to do next, reducing manual effort on tedious steps while keeping a review process in place for important decisions. The most effective AI agents are narrow, tightly scoped, and domain-specific. Here's how to approach building one. Read more

New firewall challenge metrics now available

June 3, 2025

You can now monitor and query Vercel Firewall challenge outcomes using two new metrics: challenge-solved – Visitor solved the challenge and was granted access (indicates a real user) challenge-failed – Visitor submitted an invalid challenge solution (the request was blocked) These metrics help evaluate rule effectiveness and reduce friction when users are unnecessarily challenged (high success rates). Now available in the Firewall dashboard and in the Observability Plus' query builder with no additional setup required. Learn more about custom rules and managed rulesets. Read more

Introducing the v0 composite model family

June 1, 2025

We recently launched our AI models v0-1.5-md, v0-1.5-lg, and v0-1.0-md in v0. Today, we're sharing a deep dive into the composite model architecture behind those models. They combine specialized knowledge from retrieval-augmented generation (RAG), reasoning from state-of-the-art large language models (LLMs), and error fixing from a custom streaming post-processing model. While this may sound complex, it enables v0 to achieve significantly higher quality when generating code. Further, as base models improve, we can quickly upgrade to the latest frontier model while keeping the rest of the architecture stable. Read more

Fluid compute: Evolving serverless for AI workloads

May 30, 2025

AI’s rapid evolution is reshaping the tech industry and app development. Traditional serverless computing was designed for quick, stateless web app transactions. LLM interactions require a different sustained compute and continuous execution patterns. Read more

Fluid compute now supports ISR background and on-demand revalidation

May 30, 2025

Fluid compute now supports both background and on-demand Incremental Static Regeneration (ISR) across all Vercel projects. This means ISR functions now benefit from Fluid's performance and concurrency efficiency with no config changes needed. If you’ve redeployed recently, you’re already using it. Fluid compute reuses existing resources before creating new ones, reducing costs by up to 85% for high-concurrency workloads. It delivers server-like efficiency with serverless flexibility with: Optimized concurrency Scale from zero to infinity Minimal cold starts Usage-based pricing Full Node.js and Python support No infrastructure management Background tasks with waitUntil Enable Fluid for your existing projects, and learn more in our blog and documentation. Read more

Faster login flow and new Google Sign-in support

May 30, 2025

The login experience has been redesigned for faster access and now includes full support for Google Sign-in, including Google One Tap. If your existing Vercel account's email matches your Google email, you can use the Google button from the login screen and your accounts will be automatically linked. If the emails don’t match, you can manually connect your Google account from your account settings once logged in. Read more

CVE-2025-48068

May 28, 2025

A low-severity vulnerability in the Next.js dev server has been addressed. Summary This vulnerability affects Next.js versions 13.0.0 through 14.2.29 and 15.0.0 through 15.2.1. It includes two related issues affecting the local development server: Cross-Site WebSocket Hijacking (CSWSH) and Cross-Origin Script Inclusion. Both stem from the lack of origin validation on development server resources. Impact When running next dev, a malicious website can: Initiate a WebSocket connection to localhost and interact with the local development server if the project uses the App Router, potentially exposing internal component code. Inject a tag referencing predictable paths for development scripts (e.g., /app/page.js), which are then executed in the attacker's origin. This can allow extraction of source code The root cause is insufficient origin verification on local development server resources, including the WebSocket server and static script endpoints. This issue is similar to CVE-2018-14732, though scoped strictly to local development use. Resolution This issue was fixed in Next.js versions 14.2.30 and 15.2.2. These releases introduce a configuration option to enable origin checks, which

AI query prompting now available in Observability Plus

May 28, 2025

Observability Plus users can now use natural language to create new queries or modify existing ones by adding filters, changing time ranges, or grouping results. Queries allow customers to explore log data and visualize traffic, performance, and other key metrics. AI prompts generate queries in the standard format, and are represented in the URL so they can be shared and bookmarked. Example prompts include: Show all 500 errors in the last 24 hours Show me the top bandwidth for incoming requests Show me the top hostnames grouped by country All requests challenged by DDoS mitigations by user agent Find all requests with the keyword "timeout" grouped by path This is available to all Observability Plus users at no additional cost. View the dashboard or learn more about Observability and Observability Plus. Read more

Vercel security roundup: improved bot defenses, DoS mitigations, and insights

May 23, 2025

Since February, Vercel blocked 148 billion malicious requests from 108 million unique IP addresses. Every deployment automatically inherits these protections, keeping your workloads secure by default and enabling your team to focus on shipping rather than incidents. Our real-time DDoS filtering, managed Web Application Firewall (WAF), and enhanced visibility ensure consistent, proactive security. Here's what's new since February. Read more

Middleware insights now available in Vercel Observability

May 23, 2025

The Vercel Observability dashboard now includes a dedicated view for middleware, showing invocation counts and performance metrics. Observability Plus users get additional insights and tooling: Analyze invocations by request path, matched against your middleware config Break down middleware actions by type (e.g., redirect, rewrite) View rewrite targets and frequency Query middleware invocations using the query builder View the dashboard or learn more about Observability and Observability Plus. Read more

Rate limiting now available on Hobby, with higher included usage on Pro

May 23, 2025

Rate limiting now has higher included usage and broader availability to help protect your applications from abuse and manage traffic effectively. The first 1,000,000 allowed rate limit requests per month are now included. Hobby teams also get 1 free rate limit rule per project, up to the same included allotment. These changes are now effective and have been automatically applied to your account. Learn more about configuring rate limits or create a new rate limiting rule now. Read more

Faster CDN proxying to external origins

May 23, 2025

We’ve optimized connection pooling in our CDN to reduce latency when connecting to external backends, regardless of traffic volume. Lower latency: Improved connection reuse and TLS session resumption reduce response times by up to 60% in some regions, with a 15–30% average improvement. Reduced origin load: 97% connection reuse and more efficient TLS resumption significantly cut the number of new handshakes required. This is now live across all Vercel deployments at no additional cost. Read more

External API caching insights now in Observability

May 22, 2025

The Observability dashboard now surfaces caching behavior for external API calls using Vercel Data Cache. External APIs page, you’ll see a new column indicating how many requests were served from the cache vs. the origin. Caching insights are available per hostname for all users, and per path for Observability Plus subscribers. View the external API dashboard or learn more about Vercel Data Cache. Read more

Vercel Blob is now generally available: Cost-efficient, durable storage

May 21, 2025

Storage should be simple to set up, globally available, and built to last, without slowing you down or adding complexity. It should feel native to your app. That's why we built Vercel Blob: Amazon S3-backed storage that's deeply integrated with Vercel's global application delivery and automated caching, with predictable pricing to serve public assets cost-efficiently at scale. Vercel Blob is now generally available. It's already storing and serving over 400 million files, and powers production apps like v0 and the Vercel Dashboard. Read more

Vercel Blob is now generally available

May 21, 2025

Vercel Blob is now generally available, bringing high-performance, globally scalable object storage into your workflows and apps. Blob storage’s underlying S3 infrastructure ensures 99.999999999% durability, and already stores over 400 million files while powering production apps like v0.dev. Pricing is usage-based: Storage: $0.023 per GB per month Simple API operations (e.g. Reads): $0.40 per million Advanced operations (e.g. Uploads): $5.00 per million Blob Data Transfer: starting at $0.050 per GB Pricing applies to: New Blob stores starting today Existing stores starting June 16, 2025 Hobby users now get increased free usage: 1 GB of storage and 10 GB of Blob Data Transfer per month. Get started with Vercel Blob and learn more in the documentation. Read more

How Vapi built their MCP server on Vercel

May 21, 2025

Vercel recently published a Model Context Protocol (MCP) adapter that makes it easy to spin up an MCP server on most major frameworks. Vapi is building an API for building real-time voice agents. They handle orchestration, scaling, and telephony to provide a completely model-agnostic and interchangeable interface for building agents. Vapi rebuilt their MCP server on Vercel, letting users create agents, automate testing, analyze transcripts, build workflows, and give agents access to all of Vapi’s endpoints. Read more

Introducing the AI Gateway

May 20, 2025

Note: This blog is outdated, please reference this page for the latest information or read the docs here. The Vercel AI Gateway is now available for alpha testing. Built on the AI SDK 5 alpha, the Gateway lets you switch between ~100 AI models without needing to manage API keys, rate limits, or provider accounts. The Gateway handles authentication, usage tracking, and in the future, billing. Get started with AI SDK 5 and the Gateway, or continue reading to learn more. Why we're building the AI Gateway The current speed of AI development is fast and is only getting faster. There's a new state-of-the-art model released almost every week. Frustratingly, this means developers have been locked into a specific provider or model API in their application code. We want to help developers ship fast and keep up with AI progress, without needing 10 different API keys and provider accounts. Production AI applications often run into capacity issues or rate limiting due to high demand. Infrastructure providers move quickly to bring models online and keep up with this demand, but this can come at the expense of performance or availability. The AI Gateway will allow you to load balance across provi

Vercel Blob insights now available in Observability

May 19, 2025

The Observability dashboard now includes a dedicated tab for Vercel Blob, which provides visibility into how Blob stores are used across your applications. At the team level, you can see total data transfer, download volume, cache activity, and API operations. You can also drill into activity by user agent, edge region, and client IP. This allows you to understand usage patterns, identify inefficiencies, and optimize how your application stores and serves assets. Try it out or learn more about Vercel Blob. Read more

Hypertune joins the Vercel Marketplace

May 19, 2025

Hypertune now offers a native integration with Vercel Marketplace. You can find it as a Flags & Experimentation provider in the Flags tab. The Hypertune integration offers: Powerful flags, A/B testing, analytics and app configuration Full end-to-end type safety with type-safe client generation Personalization with the lowest latency using Edge Config A first class Flags SDK adapter Install and access on Vercel with one-click setup and unified billing. Deploy the Hypertune template built for Vercel Marketplace today. Read more

How Fern delivers 6M+ monthly views and 80% faster docs with Vercel

May 15, 2025

Fern is improving how teams build and host documentation. As a multi-tenant platform, Fern enables companies like Webflow and ElevenLabs to create, customize, and serve API documentation from a single Next.js application—scaling seamlessly across multiple customer domains. With 6 million+ page views per month and 1 million+ unique visitors, performance and reliability are key. By running on Vercel’s infrastructure, Fern benefits from automatic caching, optimized content delivery, and instant scalability, all while maintaining a fast iteration cycle for development. Additionally, their migration to Next.js App Router has driven a 50-80% reduction in page load times, improving navigation speed and Lighthouse scores for customers worldwide. Read more

45% faster build initialization

May 15, 2025

Builds on Vercel now initialize 45% faster on average, reducing build times by around 15 seconds for Pro and Enterprise teams. Build initialization includes steps like restoring the build cache and fetching your code before the Build Command runs. These improvements come from continued enhancements to Hive, Vercel’s build infrastructure. This improvement also reduced I/O wait times for file writes inside the build container by 75%, improving performance for the entire build. Learn more about builds on Vercel. Read more

How Consensys rebuilt MetaMask.io with Vercel and Next.js

May 14, 2025

Since 2014, Consensys has shaped the web3 movement with tools like and Linea, Infura, and MetaMask—the most widely used self-custodial wallet on the web, with millions of users across the globe. As the blockchain ecosystem quickly matured, the need for a site that could move as fast as the teams building it became clear. To meet that demand, Consensys migrated MetaMask.io to Next.js and Vercel, creating an architecture built for scale, speed, and continuous iteration. Read more

Updated v0 pricing

May 13, 2025

We’re updating how pricing works in v0. Usage is now metered on input and output tokens which convert to credits, instead of fixed message counts. This gives you more predictable pricing as you grow and increases the amount of usage available on our free tier. Existing v0 users will transition to the new pricing at the start of your next billing period. New users will start on the improved pricing today. Read more

Proxied responses now cacheable via CDN-Cache-Control headers

May 13, 2025

Vercel’s CDN, which can proxy requests to external backends, now caches proxied responses using the CDN-Cache-Control and Vercel-CDN-Cache-Control headers. This aligns caching behavior for external backends with how Vercel Functions are already cached. This is available starting today, on all plans, at no additional cost. Per the Targeted HTTP Cache Control spec (RFC 9213), these headers support standard directives like max-age and stale-while-revalidate, enabling fine-grained control over CDN caching without affecting browser caches. You can return the headers directly from your backend, or define them in vercel.json under the headers key if your backend can't be modified. No configuration changes or redeployments required. Return the header (or set it in vercel.json) to improve performance, reduce origin load, and ensure fresh content. Learn more about CDN-Cache-Control headers. Read more

New one-click AI bot managed ruleset

May 13, 2025

You can now block AI crawlers and scrapers like GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, Bytespider (ByteDance), and others with a single toggle using the AI bot managed ruleset. Now available for free on all plans. The ruleset is managed by Vercel and updates automatically as new crawlers appear, with no additional action required. This protection operates with zero latency impact to legitimate traffic. For more complete coverage, combine with Bot Filter to catch AI bots that attempt to spoof user agents to disguise themselves as legitimate browsers or omit proper identification headers. AI crawlers now generate more traffic than human users on many popular sites, driving up infrastructure costs and raising copyright and data usage concerns. Many of these crawlers do not respect robots.txt or similar directives, making manual solutions unreliable. Enable the ruleset or learn more in the documentation. Read more

Resources tab allows instant searching and filtering of functions, middleware, and static assets

May 13, 2025

The Resources tab is replacing the Functions tab for deployments in the Vercel Dashboard, allowing you to see more than Functions. You'll now see the Resources tab when viewing a deployment in the Vercel Dashboard where you can see, search, and filter: Middleware: Any configured matchers Static Assets: Files (HTML, CSS, JS, images, fonts, and more) and their sizes Functions: The type, runtime, size, and regions You can use the three dot menu (...) to jump to the Logs, Analytics, Speed Insights, or Observability tab filtered to a given function. Read more about using the Vercel Dashboard to view and manage your deployments. Read more

The spring 2025 cohort of Vercel’s Open Source Program

May 12, 2025

Open source runs the world. The frameworks, libraries, and tools we rely on are strengthened by communities that share ideas, review code, and build in the open. At Vercel, we want to help those communities thrive. That’s why we launched the Vercel Open Source Program: a developer initiative that gives maintainers the resources, credits, and support they need to ship faster and scale confidently. Four times a year, we’ll welcome a new cohort of projects into the program. Members receive $3,600 in Vercel credits, perks from partners, and a dedicated Slack space to learn from one another. Today we are announcing this spring's cohort. Read more

New quick actions in Observability

May 9, 2025

You can now quickly copy, filter, or exclude individual results in views and query results. Copy is available across all Observability views Filter and exclude are available for custom query search results These quick actions help make it easier to explore and refine your Observability queries. Now available for Observability and Observability Plus customers. Try it out in Observability. Read more

New usage dashboard for Enterprise users

May 8, 2025

Enterprise teams with Managed Infrastructure Unit (MIU) commitments can now access a new usage dashboard with improved filtering, detailed breakdowns, and export options to better understand usage and costs by product and project. You can now break down usage by: Product to quickly identify usage, drill down into spikes, and track costs of a single or set of products Team and project to understand your costs and monitor team activity across all or specific apps CSV exports for external analysis via integration into your cost observability tools and spreadsheets Explore the new dashboard today. Read more

CDN origin timeout increased to two minutes

May 8, 2025

Vercel’s CDN will now wait up to 120 seconds for your backend to start sending data, up from 30 seconds. This extended proxied request timeout is now available on all plans at no additional cost. The proxied request timeout defines how long our CDN allows your external backend to respond before canceling the request. After the initial byte is received, your backend can take longer than two minutes to complete the request, as long as it continues sending data at least once every 120 seconds. This update improves reliability for workloads with long processing times, such as LLM generation or complex data queries, and reduces the chance of 504 gateway timeouts. This change is effective immediately, with no action or configuration required. Read more

Up to 80% pricing reduction for Web Analytics

May 8, 2025

We’ve increased included limits and reduced the price of Web Analytics event and the Web Analytics Plus add-on by up to 80%. Web Analytics is now billed: Per single event, instead of 100K increments At $0.00003 per event ($3 per 100K, a 79% decrease from news: [4 per 100K) At news: [0/month for the Plus add-on (an 80% decrease, from $50/month) Web Analytics Plus is an optional add-on that unlocks increased retention and UTM parameters. Included event limits have increased: Hobby: 50K events/month (20x increase, from 2.5K) Pro: 100K events/month (4x increase, from 25K) Learn more about Web Analytics pricing. Read more

Flags Explorer is now generally available

May 7, 2025

The Flags Explorer lets you override feature flags for your own session, without affecting colleagues, and without signing into your flag provider. This enables you to test features in production before they go live and keeps you in the flow. This feature is now generally available for all customers. Hobby, Pro, and Enterprise plans include 150 overrides per month, with unlimited overrides available for $250 per month on Pro and Enterprise. Teams that used Flags Explorer during the beta have 30 days to activate the new unlimited option before the 150 overrides per month limit takes effect. This can be done in the Vercel dashboard or directly through the Vercel Toolbar. Additionally, The Flags SDK automatically respects overrides set by the Flags Explorer, no matter which adapter you're using. Learn more about Flags Explorer. Read more

Faster builds now available with compute upgrades on paid plans

May 7, 2025

Projects with on-demand concurrent builds can now use enhanced build machines to improve build performance. Available on all paid plans, these machines offer double the resources: 8 CPUs, 16 GB memory, and 58 GB disk. This reduces both build time and total build minutes used. Existing customers are already seeing up to 25% faster builds with no changes required. Enhanced builds can be enabled per project and are billed per minute. Enterprise customers can run all concurrent builds, including pre-allocated build slots and on-demand, on higher-spec machines. Enable on-demand enhanced builds and learn more in our documentation. Read more

Introducing the Flags Explorer, first-party integrations, and updates to the Flags SDK

May 7, 2025

Experimentation, A/B testing, and feature flags serve as essential tools for delivering better user experiences, ensuring smoother rollouts, and empowering teams to iterate quickly with confidence. We're making it easier to bring flags into your workflow with: Read more

MCP server support on Vercel

May 7, 2025

Model Context Protocol (MCP) is a way to build integrations for AI models. Vercel now supports deploying MCP servers (which AI models can connect to) as well as MCP clients (AI chatbot applications which call the servers). Get started with our Next.js MCP template today. How is MCP different than APIs? APIs allow different services to communicate together. MCP is slightly different. Rather than thinking about MCP like a REST API, you can instead think about it like a tailored toolkit that helps an AI achieve a particular task. There may be multiple APIs and other business logic used behind the scenes for a single MCP tool. If you are already familiar with tool-calling in AI, MCP is a way to invoke tools hosted on a different server. MCP now supports a protocol similar to other web APIs, namely using HTTP and OAuth. This is an improvement from the previous stateful Server-Sent Events (SSE) protocol. Deploying MCP servers to Vercel To simplify building MCP servers on Vercel, we’ve published a new package, @vercel/mcp-adapter, which supports both the older SSE transport and the newer stateless HTTP transport. The majority of MCP clients currently only support the SSE transport option.

Bot activity and crawler insights now in Observability

May 7, 2025

Vercel Observability now provides detailed breakdowns for individual bots and bot categories, including AI crawlers and search engines. Users across all plans can view this data in the Observability > Edge Requests dashboard. Additionally, Observability Plus users can: Filter traffic by bot category, such as AI View metrics for individual bots Break down traffic by bot or category in the query builder Inspect bot and crawler activity in your Observability dashboard now. Read more

Join the Vercel AI Accelerator

May 6, 2025

The Vercel AI Accelerator is back. This year, we'll work with 40 teams building the future of AI. Over six weeks, participants get the tools, infrastructure, and support to create next-generation AI apps. Applications are open now until May 17. Read more

Track a request's full lifecycle with session tracing

May 5, 2025

Session tracing is now available to all Vercel users, providing end-to-end visibility into the timing of each step in a request's lifecycle, from when it enters Vercel’s infrastructure to execution inside your Vercel Functions. With session tracing you can: Start tracing sessions on your deployments directly from the Vercel Toolbar, no setup required. View spans for Vercel's routing, caching, middleware, and function layers as well as those instrumented in your code. Share traces with teammates for faster debugging and optimization. Use tracing alongside logs and metrics to debug, optimize, and improve iteration speed. Session tracing is free to customers on all plans. To get started, find Tracing in the Vercel Toolbar, or learn more in the docs. Read more

How v0 is building SEO-optimized sites by default

May 2, 2025

Building for the web goes beyond speed and aesthetics, discoverability matters just as much. While AI can accelerate web development, it often skips over performance, accessibility, or SEO best practices that matter for discoverability. With v0, you don’t have to compromise. Every interface you generate is fast, accessible, and SEO-optimized by default. v0 integrates with Next.js and deploys to Vercel, giving you structured metadata, performance tuning, and Server Side Rendering (SSR). The result is better Core Web Vitals, pages that load quickly and return full HTML, making them easier for search engines to crawl and index. Read more

Information disclosure in Flags SDK (CVE-2025-46332)

May 2, 2025

Vercel discovered and patched an information disclosure vulnerability in the Flags SDK, affecting versions: flags ≤ 3.2.0 @vercel/flags ≤ 3.1.1 This is being tracked as CVE-2025-46332. We have published an automatic mitigation for the default configuration of the Flags SDK on Vercel. We recommend upgrading to flags@4.0.0 (or migrating from @vercel/flags to flags) to remediate the issue. Further guidance can be found in the upgrade guide. Impact and analysis A malicious actor could determine the following under specific conditions: Flag names Flag descriptions Available options and their labels (e.g. true, false) Default flag values Flags providers were not accessible. No write access nor additional customer data was exposed, this is limited to the values noted above. Automatic mitigation Vercel implemented a network-level mitigation to prevent the default flags discovery endpoint at /.well-known/vercel/flags being reachable, which automatically protects Vercel deployments against exploitation of this issue. While uncommon, if you are exposing the flags discovery endpoint through custom paths, you can also implement a custom WAF rule to restrict access to these endpoints as a mitiga

Serve personalized content faster with Vary support

May 2, 2025

Vercel now fully supports the HTTP Vary header, making it easier to cache personalized content across all plans with no configuration required. The Vary header tells caches which request headers to include when generating cache keys. This allows Vercel’s application delivery network to store and serve different versions of a page based on headers like X-Vercel-IP-Country or Accept-Language, so users get fast, localized content without recomputation. By returning the above headers your site caches and serves country-specific content. A visitor from the United States receives the US-specific cached version, and visitors from other countries receive the version for their locale, with no recomputation required. Learn more about caching personalized content in Vercel's application network documentation. Read more

iOS developers can now offer commission-free payments on web

May 1, 2025

Yesterday, a federal court made a decisive ruling in Epic Games v. Apple: Apple violated a 2021 injunction by continuing to restrict developers from linking to external payment methods, and by imposing a 27% fee when they did. The ruling represents a major shift for native app developers. Read more

Create custom WAF rules directly from the Vercel Firewall tab

May 1, 2025

You can now create custom WAF rules directly from the chart displayed on the Firewall tab of the Vercel dashboard. When viewing your traffic grouped by a parameter (like IP address, user agent, or request path), you can now select "Create Custom Rule" within the actions menu of any displayed time series. This automatically generates an editable draft of the custom WAF rule that matches the selected parameter. Once the WAF rule is saved and published, it's immediately propagated across our global network. This feature is available to all users across all plans at no additional cost. Learn more about the Vercel Firewall. Read more

Optionally disable deployment_status webhook events for GitHub Actions

May 1, 2025

You can now disable the deployment_status webhook event that Vercel sends to GitHub when Vercel is connected to your GitHub repository. When deployment_status events are enabled, GitHub's pull request activity will create a log with a status event for every deployment. While this can keep your team better informed, it can also create noisy event logs for repositories with many deployment events, especially in monorepos with many projects. Disabling these events prevents repeated messages from cluttering your GitHub PR's event history, giving you a cleaner, more focused view of your pull request activity. The Vercel Github comment containing links to your preview deployments will continue to be posted as before. The deployment_status event is most often used as a trigger for GitHub Actions. We recommend migrating to repository_dispatch events to simply workflows with richer Vercel deployment information. Learn more in the documentation. Read more

Checks API support added for Marketplace integration providers

April 30, 2025

Providers building native integrations for the Vercel Marketplace can now use the Checks API to deliver deeper functionality for their users. With Vercel's Checks API, you can define and run custom tests and assertions after every deployment, then surface actionable results directly in the Vercel dashboard. As a testing provider, you can implement checks such as reliability tests (e.g. API availability, runtime errors), performance tests (e.g. response time thresholds, load simulation), or Web Vitals (e.g. layout shift). This helps developers catch real-world issues earlier in their workflow, powered by your integration. When building your integration, keep these best practices in mind: Offer minimal or no-configuration solutions so developers can easily run checks Provide a guided onboarding experience from installation to first results Display clear, actionable outcomes directly in the Vercel dashboard Document ways to extend or customize checks for advanced users Learn more in the Checks API documentation. Read more

Protection against React Router and Remix vulnerabilities

April 26, 2025

Security researchers reviewing the Remix web framework have discovered two high-severity vulnerabilities in React Router. Vercel proactively deployed mitigation to the Vercel Firewall and Vercel customers are protected. CVE-2025-43864 and CVE-2025-43865 enable an external party to modify the response using certain request headers, which can lead to cache poisoning Denial of Service (DoS). CVE 43865 enables vulnerabilities such as stored Cross Site Scripting (XSS). Impact and analysis When we learned about the vulnerability, we started analyzing the impact to the Vercel platform. Here are our findings and recommendations: We were able to reproduce the vulnerability and demonstrate that cache poisoning is trivial, including stored Cross Site Scripting (XSS) injections The only precondition is that the customer used an impacted version of Remix / React Router (v7.0.0 branch prior to version v7.5.2) and Cache-Control headers The impact can extend to any visitor of the application after the cache is poisoned, regardless of authentication state or any other request headers Vercel customers using React Router between v7.0.0 and v7.5.1 were impacted before our Firewall mitigation We have d

Improved experience for managing project domains

April 25, 2025

We’ve redesigned the Project Domains page with faster search, smoother navigation, and clearer visibility into your domain configurations. Faster Browsing and Cleaner Overviews Navigating and understanding your domain setup is now quicker and more direct: Live Search: Start typing in the search bar, and your domain list will filter as you type without needing an exact match. Infinite Scroll: We've replaced the "View More" button with smooth, infinite scrolling so you can browse without interruptions. Cleaner View: Key information like associated Redirects and Environments are now displayed inline within the domain list, giving you a comprehensive overview at a glance without needing to click into individual domain details. Streamlined Configuration and Setup Configuring DNS and adding new domains is now more focused and user-friendly: Focused DNS Configuration in Modals: We’ve moved the DNS configuration instructions into a modal. This allows you to focus solely on configuring the domain you’ve added. Guided Full-Page Add Flow: Adding a new domain is now a clearer, step-by-step process with our new full-page add flow. We guide you through the necessary configurations to ensure a co

Pro customers can now deploy faster without build queues

April 25, 2025

When multiple team members deploy to Vercel at once, builds by default have a queue. Now, you can remove these queues, enabling your builds to start immediately. This is available to both Pro and Enterprise customers with new per-minute pricing and can be applied in the following ways: Manually, per deployment, for urgent builds Automatically, at the project level, to avoid queues by default Concurrent build slots remain available for teams with steady, high-volume workloads. Learn about on-demand concurrent builds and enable them on your project. Read more

Pricing for on-demand concurrent builds reduced by over 50%

April 25, 2025

Pricing for on-demand concurrent builds, which allow deployments to bypass build queues, has been reduced by more than 50%. Usage increments have also been lowered from 10 minutes to 1 minute. On-demand concurrent builds are available to both Pro and Enterprise customers, and complement existing build slots with the following recommendations: Use on-demand for bursty workloads or priority deploys Use slots for large, frequent builds with predictable volume This change also applies to all customers using Enhanced On-demand builds, which allocate more memory to build compute for faster deployment times. Learn about on-demand concurrent builds and enable them on your project. Read more

Updates to Vercel Toolbar shortcuts

April 24, 2025

You can now customize keyboard shortcuts for the Vercel Toolbar. Replace default shortcuts for hiding and opening the Toolbar Menu, and add shortcuts for frequently used tools. To configure shortcuts, find Keyboard Shortcuts under Preferences in the Toolbar Menu. The browser extension is needed to customize shortcuts for hiding the toolbar and opening the Toolbar Menu. The default shortcut to show and hide the Toolbar Menu is changing to reduce conflicts with sites that have their own Cmd+K menus. Mac: changing from ⌘ K to ^ (control) Windows: changing from Ctrl K to Ctrl Learn more about the Vercel Toolbar. Read more

Prisma joins the Vercel Marketplace

April 23, 2025

Prisma is now available as a storage provider on the Vercel Marketplace, offering Prisma Postgres, a serverless database optimized for fullstack and edge applications. With automated account creation, integrated billing through Vercel, and a generous free tier, developers can now get started with Prisma Postgres in just a few clicks, no separate signup required. With the Prisma native integration, Vercel users get: A high-performance Postgres database with zero cold starts Automatic scaling with built-in global caching and connection pooling Visual data management and AI-powered performance suggestions Get started with Prisma on the Vercel Marketplace. Available to customers on all plans. Read more

Node.js Vercel Functions now support request cancellation

April 23, 2025

Vercel Functions using Node.js can now detect when a request is cancelled and stop execution before completion. This includes actions like navigating away, closing a tab, or hitting stop on an AI chat to terminate compute processing early. This reduces unnecessary compute, token generation, and sending data the user never see. You can listen for cancellation using Request.signal.aborted or the abort event: If you're using the AI SDK, forward the abortSignal to your stream: Learn more about cancelling Function requests. Read more

Fluid compute is now the default for new projects

April 23, 2025

New Vercel projects now run on Fluid compute by default. This update follows Fluid’s general availability, its adoption across large-scale production apps, and all v0.dev deployments shipping with Fluid enabled by default. Fluid compute reuses existing instances before spawning new ones, cutting costs by up to 85% for high-concurrency workloads. It combines the efficiency of servers with the flexibility of serverless: Concurrent requests per function Scale from zero to infinity Minimal cold starts Usage-based, pay as you go Full Node.js and Python support No infrastructure to manage Background tasks with waitUntil Enable Fluid for your existing projects, and learn more in our blog and documentation. Read more

Bot Protection: One-click managed ruleset now in public beta

April 23, 2025

The Vercel Web Application Firewall (WAF) inspects billions of requests every day to block application-layer threats, such as cross-site scripting, traversal, and application DDoS attacks. While we already inspect and block malicious bot traffic, we wanted to provide better, more precise controls to fine tune your application security. Today, we're launching the Bot Protection managed ruleset, free for all users on all plans. With a single click, you can protect your application from bot attacks. Read more

Bot Protection is now in public beta

April 23, 2025

Vercel Web Application Firewall now includes a new Bot Protection managed ruleset, available in public beta for all users. Bot Protection helps reduce automated traffic from non-browser sources and allows you to respond based on two action choices: Log Only Action: Logs identified bot traffic in the Firewall tab without blocking requests Challenge Action: Serves a browser challenge to traffic from non-browser sources. Verified bots are automatically excluded To avoid disrupting legitimate automated traffic that's not already covered by Verified Bots, you can configure custom WAF rules using the bypass action for specific requests. To enable the ruleset: In your project dashboard, navigate to the Firewall tab and select Configure Under Bot Management, navigate to Bot Protection Select Log or Challenge Select Review Changes and review the changes to be applied Select Publish to apply the changes to your production deployment Bot Protection complements Vercel's existing mitigations, which already block common threats like DDoS attacks, low quality traffic, and spoofed traffic. It adds an extra layer of protection for any automated traffic that is not clearly malicious. During this pub

CVE-2025-32421

April 22, 2025

A low severity cache poisoning vulnerability was discovered in Next.js. Summary This affects versions >14.2.24 through 14.2.24 through <15.1.6 Impact This vulnerability allows an attacker to poison the CDN cache by injecting the response body from a non-cacheable data request (?__nextDataRequest=1) into a normal request that retains cacheable headers, such as Cache-Control: public, max-age=300. No backend access or privileged escalation is possible through this vulnerability. This issue was verified using automated tooling that repeatedly triggers the race condition. Successful exploitation depends on precise timing and the presence of a vulnerable CDN configuration. A Python-based proof of concept script was shared by the reporter and used to validate this behavior on live targets prior to the patch. Patches This issue was patched in 15.1.6 and 14.2.24 by stripping the x-now-route-matches header from incoming requests. Vercel Platform Mitigation Applications hosted on Vercel's platform are not affected by this issue, as the platform does not cache responses based solely on 200 OK status without explicit cache-control headers. Workarounds For self-hosted Next.js deployments unable

Becoming an AI engineering company

April 18, 2025

In today's rapidly evolving tech landscape, AI has moved from research labs to everyday tools with stunning speed. I wanted to share my perspective, not only as a CTO at Vercel, but as an engineer who's seen a few revolutions over the past 30 years. Read more

Lower pricing for Fast Data Transfer

April 17, 2025

Today we are lowering the price of Fast Data Transfer (FDT) for Vercel regions in Asia Pacific, Latin America, and Africa by up to 50%. The new FDT regional pricing is rolling out for all Pro and Enterprise plans: All new Pro and Enterprise users will be charged the new price moving forward. For existing Pro users, the new pricing applies starting today. For existing Enterprise users, it will apply at the start of the next billing cycle (typically monthly). Vercel Region Old price per GB New price per GB Cape Town, South Africa (cpt1) $0.39 $0.28 Hong Kong (hkg1) $0.30 $0.16 Mumbai, India (bom1) $0.33 $0.20 Osaka, Japan (kix1) $0.31 $0.16 Sao Paulo, Brazil (gru1) $0.44 $0.22 Seoul, South Korea (icn1) $0.47 $0.35 Singapore (sin1) $0.30 $0.16 Sydney, Australia (syd1) $0.32 $0.16 Tokyo, Japan (hnd1) $0.31 $0.16 Learn more about Fast Data Transfer or review your FDT usage on the Usage page. Read more

Enhanced Builds now have double the compute

April 17, 2025

Enhanced Builds now offer double the compute capacity, further improving performance for large codebases and CPU-intensive builds. Available to Enterprise customers, Enhanced Builds are designed for teams working with monorepos or frameworks that run tasks in parallel—like dependency resolution, transpilation, or static generation. Customers already using Enhanced Builds are seeing, with no action required, up to 25% reductions in build times. Learn more in our documentation or speak to your Vercel account team to enable Enhanced Builds. Read more

Protection against React Router vulnerability CVE-2025-31137

April 17, 2025

Security researchers reviewing the Remix web framework have recently discovered a high-severity vulnerability in React Router that allows URL manipulation through the Host/ X-Forwarded-Host header. Our investigation determined that Vercel and our customers are unaffected: We use query parameters as part of the cache key, which protects against cache poisoning driven by the _data query praram. The @vercel/remix adapter uses X-Forwarded-Host similarly to the Express adapter, but it is not possible for an end user to send X-Forwarded-Host to a Function hosted on Vercel. A patch has been issued and released in Remix 2.16.3 / React Router 7.4.1. We recommend customers update to the latest version. Read more about CVE-2025-31137. Read more

Life of a Vercel request: Application-aware routing

April 15, 2025

Routing is a fundamental part of delivering applications, but it’s often treated as an afterthought—tacked onto the caching layer and configured through complex YAML or manual click-ops. This can introduce friction for teams, increase the risk of misconfigurations, and slow down deployments, especially as applications grow in complexity. Vercel takes a different approach: routing is built into the platform as an application-aware gateway that understands your codebase. This unlocks a range of capabilities that simplify development by reducing configuration overhead, minimizing latency, and enabling more advanced architectures. The gateway has full context of your deployments, domains, and logic. It supports standard routing and custom rules, but goes beyond reverse proxying by interpreting application logic in real time to make smarter decisions, like skipping unnecessary compute. Here’s how Vercel routes requests—and why it makes building performant, complex apps easier. Read more

Update on Spain and LALIGA blocks of the internet

April 15, 2025

A Spanish court has granted LALIGA the power to block IP addresses associated with unauthorized football streaming—without distinguishing between infringing and non-infringing services. As a result, legitimate, unrelated websites that people depend on are now inaccessible in Spain. Read more

Migrating Grep from Create React App to Next.js

April 14, 2025

Grep is extremely fast code search. You can search over a million repositories for specific code snippets, files, or paths. Search results need to appear instantly without loading spinners. Originally built with Create React App (CRA) as a fully client-rendered Single-Page App (SPA), Grep was fast—but with CRA now deprecated, we wanted to update the codebase to make it even faster and easier to maintain going forward. Here's how we migrated Grep to Next.js—keeping the interactivity of a SPA, but with the performance improvements from React Server Components. Read more

Vercel Observability is now route-aware for SvelteKit apps

April 14, 2025

SvelteKit routes with dynamic segments—like /blog/[slug]—are now individually recognized and surfaced by Vercel Observability. This replaces the previous behavior where all dynamic routes appeared under a single /fn entry. This is available with version 5.7.0 of @sveltejs/adapter-vercel. Upgrade to unlock improved observability for your SvelteKit projects. Learn more about Vercel Observability. Read more

Legacy build image is being deprecated on September 1, 2025

April 10, 2025

Node.js 18 (LTS support ends April 30, 2025) and the Vercel legacy build image will be deprecated on September 1, 2025. If you are still using the legacy build image on this date, new builds will display an error. What changes between the legacy build image and latest build image? The minimum version of Node.js is now 20.x The Python toolchain version is now 3.12 The Ruby toolchain version is now 3.3.x How do I know if I am still using the legacy build image? Projects using Node.js 18.x in Build and Deployment Settings use the legacy build image Projects using overrides in package.json use the legacy build image Will my existing deployments be affected? Existing deployments will not be affected. However, the Node.js version will need to be updated on your next deployment. How can I see if my projects are affected? You can see which projects are affected by this deprecation by running the following commands: How do I upgrade? To upgrade with the dashboard, visit the Build and Deployment settings for your project and upgrade the version. To upgrade with code, use the engines field in package.json: This date coincides with the previously announced deprecation of Node.js 18 on the Verc

Introducing Chatbot Template

April 9, 2025

Update: Chat SDK has been renamed to Chatbot template, and a new Chat SDK is now available to provide a unified language for chat bots across Slack, Teams, GitHub, and Discord. The AI SDK powers incredible applications across the web, and today we're announcing the Chatbot—a best-in-class, production-ready template for building conversational AI applications like ChatGPT or Claude artifacts. Read more

Grok 3 now available on Vercel Marketplace

April 9, 2025

xAI's latest and most powerful Grok 3 models are now available through the Vercel Marketplace, bringing state-of-the-art AI capabilities to your Vercel projects. To get started, you can use the AI SDK xAI provider in your project: Then, install the xAI Marketplace Integration with Vercel CLI (or from the dashboard): Once you've accepted the terms, you'll be able to use Grok models from within your project, with no additional steps necessary. To help you get started, we've also made a ready-to-deploy Next.js xAI starter template. To learn more about xAI on Vercel, read our announcement and the documentation. Read more

Expanding observability on Vercel

April 8, 2025

The Vercel Marketplace adds new integrations from Sentry, Checkly, and Dash0. You can now use the tools you already trust to monitor, measure, and debug your apps. No custom setup. No change to how you build or deploy. These tools connect directly through the Vercel Marketplace with integrated billing, single sign-on, and access to provider dashboards, giving you deep visibility without the setup overhead. Read more

Automatic mitigation of Google and Bing crawl delay, via Vercel’s Skew Protection

April 8, 2025

Google and Bing web crawlers occasionally crawl a document, but render it up to several weeks later using a headless browser. This delay between document crawl and assets download (which happens during render) can cause indexing failures if the website has been re-deployed since the crawl. Vercel now automatically protects against such indexing failures for projects that have Skew Protection enabled. This was achieved by extending the maximum age for Skew Protection to 60 days for requests coming from major search engine bots, such as Googlebot and Bingbot. This means that assets deployed up to 60 days ago will still be accessible to search engines when they render your document. Regardless of the maximum age configured in the dashboard, Pro and Enterprise accounts using Skew Protection will automatically be protected from this delay, thereby improving SEO. Learn more about Skew Protection and enable it in your project. Also, check out our SEO research on how Google handles JavaScript throughout the indexing process, which provides a deeper dive into the search rendering process. Read more

Sentry, Checkly, and Dash0 join the Vercel Marketplace

April 8, 2025

New native integrations from Sentry, Checkly, and Dash0 are now available on the Vercel Marketplace, helping make it easier to monitor, debug, and optimize your applications—all in one place. Sentry: Real-time error tracking and performance monitoring for faster issue resolution Checkly: End-to-end monitoring and synthetic checks for your frontend and APIs Dash0: Log management and structured observability, built with a developer-first experience. Dash0 also supports Native Log Drains, allowing you to stream logs from your Vercel projects to external logging systems for deeper insights and centralized monitoring This launch introduces Log Drain support for native integrations—a capability that was previously only available to connectable accounts. These integrations offer frictionless onboarding, single sign-on, and integrated billing through Vercel, making it easy to get started in just a few clicks. Explore the new observability integrations. Read more

Protectd: Evolving Vercel’s always-on denial-of-service mitigations

April 7, 2025

Securing web applications is core to the Vercel platform. It’s built into every request, every deployment, every layer of our infrastructure. Our always-on Denial-of-Service (DoS) mitigations have long run by default—silently blocking attacks before they ever reach your applications. Last year, we made those always-on mitigations visible with the release of the Vercel Firewall, which allows you to inspect traffic, apply custom rules, and understand how the platform defends your deployments. Now, we’re introducing Protectd, our next-generation real-time security engine. Running across all deployments, Protectd reduces mitigation times for novel DoS attacks by over tenfold, delivering faster, more adaptive protection against emerging threats. Let's take a closer look at how Protectd extends the Vercel Firewall by continuously mapping complex relationships between traffic attributes, analyzing, and learning from patterns to predict and block attacks. Read more

Trigger GitHub Actions with enriched deployment data from Vercel

April 7, 2025

You can now trigger GitHub Actions workflows in response to Vercel deployment events with enriched data using repository_dispatch events. These events are sent from Vercel to GitHub, enabling more flexible, cost-efficient CI workflows, and easier end-to-end testing for Vercel deployments. Previously, we recommended using deployment_status events, but these payloads were limited and required extra parsing or investigation to understand what changed. With repository_dispatch, Vercel sends custom JSON payloads with full deployment context—allowing you to reduce Github Actions overhead and streamline your CI pipelines. We recommend migrating to repository_dispatch for a better experience. deployment_status events will continue to work for backwards compatibility. Read more

Llama 4 is now available on Vercel Marketplace

April 5, 2025

Meta’s latest and most powerful Llama 4 models are now available through the Vercel Marketplace via Groq. To get started for free, install the Groq integration in the Vercel dashboard or add Groq to your existing projects with the Vercel CLI: You can then use the AI SDK Groq provider with Lama 4: For a full demo, check out the official Groq chatbot template (which now uses Llama 4) or compare Llama 4 against other models side-by-side on our AI SDK Playground. To learn more, visit our AI documentation. Read more

Run and share custom queries in Observability Plus

April 4, 2025

Observability Plus customers can now create and share custom queries directly from the Observability dashboard—making it easier to investigate specific metrics, routes, and application behavior without writing code. The new query interface lets you: Filter by route to focus on specific pages and metrics Use advanced filtering, with auto-complete—no query language needed Analyze charts in the context of routes and projects Share queries instantly via URL or Copy button This new querying experience builds on the Monitoring dashboard, helping you stay in context as you drill deeper into your data. To try it out, open your Observability dashboard and select Explore query arrows on any chart or the query builder from the ellipsis menu. Learn more about running queries in Observability and its available metrics. Read more

Two-Factor Authentication (2FA) is now available

April 3, 2025

Users can now secure their accounts using Two-Factor Authentication (2FA) with Time-based One-Time Passwords (TOTP), commonly provided by authenticator apps like Google Authenticator or Authy. Your current Passkeys (WebAuthn keys) can also be used as second factors. 2FA adds an extra security layer to protect your account even if the initial login method is compromised. To Enable 2FA: Navigate to Authentication in Account Settings and enable 2FA Log in using your existing method (email OTP or Git provider) as your first factor Complete authentication with a TOTP authenticator as your second factor Important information: Passkey logins (WebAuthn) are inherently two-factor and won't prompt for additional verification Team-scoped SAML SSO logins delegate authentication responsibility to your identity provider (IdP) and won't require an additional factor within Vercel Visit your account settings to enable 2FA today, or check out our documentation to learn more. Read more

How PAIGE grew revenue by 22% with Shopify, Next.js, and Vercel

April 3, 2025

PAIGE, a leading denim and apparel retailer, faced significant technical complexity due to their existing ecommerce architecture. Seeking a faster and more reliable online experience, they reimagined their ecommerce strategy by adopting a simpler headless tech stack—one powered by Shopify, Next.js, and Vercel—that ultimately boosted their Black Friday revenue by 22% and increased conversion rates by 76%. Read more

Vercel Secure Compute now supports multiple environments

April 3, 2025

Teams using Vercel Secure Compute can now associate each project environment—Production, Preview, and custom—with a distinct Secure Compute network, directly from the project settings. This simplifies environment-specific network isolation within a single project. To connect your project's environments to Secure Compute: Navigate to your project's Secure Compute settings For every environment you want to connect to Secure Compute: Select an active network Optionally, select a passive network to enable failover Optionally, enable builds to include the project's build container in the network Click Save to persist your changes Learn more about Secure Compute. Read more

CVE-2025-30218

April 2, 2025

In the process of remediating CVE-2025-29927, we looked at other possible exploits of Middleware. We independently verified this low severity vulnerability in parallel with two reports from independent researchers. Summary To mitigate CVE-2025-29927, Next.js validated the x-middleware-subrequest-id which persisted across multiple incoming requests: However, this subrequest ID is sent to all requests, even if the destination is not the same host as the Next.js application. Initiating a fetch request to a third-party within Middleware will send the x-middleware-subrequest-id to that third party. Impact While the exploitation of this vulnerability is unlikely due to an attacker requiring control of the third-party, we want to be proactive. We were already planning on removing this recursion prevention logic from Middleware—it was not supported in newer updates to Middleware to support the Node.js runtime—this disclosure expedited our efforts to bring parity between runtimes. Vercel customers are protected with mitigations already implemented within our platform environment. We still encourage teams to update to the latest Next.js patch version or their chosen backport. Other infrastru

The no-nonsense guide to composable commerce

April 1, 2025

Composable commerce projects frequently become overly complex, leading to missed objectives and unnecessary costs. At Vercel, we take a no-nonsense approach to composable commerce that's solely focused on business outcomes. Architecture should serve the business, not the other way around. Ivory tower architectures disconnected from clear business goals inevitably lead to projects plagued by runaway costs. Here are five truths we stand by when it comes to composable commerce: Read more

Attack Challenge Mode now allows verified bots and Vercel cron jobs

April 1, 2025

Verified webhook providers—including Stripe and PayPal—are now automatically allowed in Attack Challenge Mode, ensuring uninterrupted payment processing. Well-behaved bots from major search engines, such as Googlebot, and analytics platforms are also supported. Vercel Cron Jobs are now exempt from challenges when running in the same account. Like other trusted internal traffic, they bypass Attack Challenge Mode automatically. To block specific known bots, create a custom rule that matches their User Agent. Known bots are validated to be authentic and cannot be spoofed to bypass Attack Challenge Mode. Learn more about Attack Challenge Mode and how Vercel maintains its directory of legitimate bots. Read more

Yarn 2+ dependency caching now supported

March 31, 2025

Vercel now caches dependencies for projects using Yarn 2 and newer, reducing install times and improving build performance. Previously, caching was only supported for npm, pnpm, Bun, and Yarn 1. To disable caching, set the environment variable VERCEL_FORCE_NO_BUILD_CACHE with a value of 1 in your project settings. If you're using Yarn 4, enable Corepack, as recommended by Yarn. Visit the Build Cache documentation to learn more. Read more

Flags SDK 3.2

March 31, 2025

The Flags SDK 3.2 release adds support for precomputed feature flags in SvelteKit, making it easier to experiment on marketing pages while keeping them fast and avoiding layout shift. Precomputed flags evaluate in Edge Middleware to decide which variant of a page to show. This keeps pages static, resulting in low global latency as static variants can be served through the Edge Network. Precompute handles the combinatory explosion when using multiple feature flags statically. Generate different variants of a page at build time, rely on Incremental Static Regeneration to only build a specific combinations on demand, and more. We also improved the Flags SDK documentation by splitting it across different frameworks and explicitly listing all providers that have adapters for the Flags SDK. Learn more about the Flags SDK with SvelteKit and the precompute pattern. Read more

Custom reporting now available on AI Gateway

March 25, 2025

Custom Reporting for AI Gateway is now available in beta for Pro and Enterprise users. Slice your AI spend by any dimension: model, provider, user, custom tags, or credential type. A single place for all your AI cost and usage data, including BYOK. All your inference already runs through AI Gateway, which means cost, token, and request data across every model, provider, and credential type is already there. The Custom Reporting API makes it queryable, so you're not logging into separate provider consoles or maintaining a third-party proxy for cost tracking. How it works Tag your requests with `user` and `tags` to attribute costs by customer, feature, team, or environment: Tagging works with the AI SDK, Chat Completions API, Responses API, OpenResponses API, and Anthropic Messages API. Then query the reporting endpoint: Group by time, tag, user, model, provider, and all the other AI parameters. Example Use the custom reporting endpoint to get all your customer costs by use case. Pricing Charge type Cost Tag/User ID $0.075 per 1,000 unique tag or user ID values written per request Query $5 per 1,000 queries to the reporting endpoint To learn more, read the AI Gateway custom reporting

Postmortem on Next.js Middleware bypass

March 25, 2025

Last week, we published CVE-2025-29927 and patched a critical severity vulnerability in Next.js. Here’s our post-incident analysis and next steps. Read more

Protection against Next.js CVE-2025-29927

March 22, 2025

A security vulnerability in Next.js was responsibly disclosed, which allows malicious actors to bypass authorization in Middleware when targeting the x-middleware-subrequest header. Vercel customers are not affected. We still recommend updating to the patched versions. Learn more about CVE-2025-29927. Read more

AI SDK 4.2

March 21, 2025

The AI SDK is an open-source toolkit for building AI applications with JavaScript and TypeScript. Its unified provider API allows you to use any language model and enables powerful UI integrations into leading web frameworks such as Next.js and Svelte. Read more

Flags SDK now supports OpenFeature

March 21, 2025

The Flags SDK adapter for OpenFeature allows using any Node.js OpenFeature provider with the Flags SDK. Pick from a wide range of flag providers, while benefiting from the Flag SDK's tight integration into Next.js and SvelteKit. OpenFeature is an open specification that provides a vendor-agnostic, community-driven API for feature flagging that works with your favorite feature flag management tool or in-house solution. OpenFeature exposes various providers through a unified API. The Flags SDK sits between your application and the source of your flags, helping you follow best practices and keep your website fast. Use the Flags SDK OpenFeature adapter in your application to load feature flags from all compatible Node.js OpenFeature providers, including: AB Tasty Bucket Cloudbees Confidence by Spotify ConfigCat DevCycle Environment Variables Provider FeatBit flagd Flipt GO Feature Flag GrowthBook Hypertune Kameleoon LaunchDarkly PostHog Split View the OpenFeature adapter or clone the template to get started. Read more

xAI and Vercel partner to bring zero-friction AI to developers

March 20, 2025

Vercel provides the tools and infrastructure to build AI-native web applications. We're partnering with xAI to bring their powerful Grok models directly to Vercel projects through the Vercel Marketplace—and soon v0—with no additional signup required. To help you get started, xAI is introducing a new free tier through Vercel to enable quick prototyping and experimentation. These Grok models now power our official Next.js AI chatbot template with the AI SDK. This is a part of our ongoing effort to make using AI frictionless on Vercel. Read more

xAI joins the Vercel Marketplace

March 20, 2025

xAI's Grok models are now available in the Vercel Marketplace, making it easy to integrate conversational AI into your Vercel projects. Get started with xAI's free plan—no additional signup through the Marketplace Access Grok's large language models (LLMs) directly from your Vercel projects Simplify authentication and API key management through automatically configured environment variables Pay only for what you use with integrated billing through Vercel To get started, you can use the AI SDK xAI provider in your project: Then, install the xAI Marketplace Integration with Vercel CLI (or from the dashboard): Once you've accepted the terms, you'll be able to use Grok models from within your project, with no additional steps necessary. To help you get started, we've also made a ready-to-deploy Next.js xAI starter template. To learn more about xAI on Vercel, read our announcement and the documentation. Read more

Lockfile-aware deployment skipping for monorepos

March 20, 2025

Vercel now maps dependencies in your package manager’s lockfile to applications in your monorepo. Deployments only occur for applications using updated dependencies. This feature is based on Turborepo's lockfile analysis, supporting the package managers listed as stable in Turborepo's Support Policy. Previously, any change to the lockfile would redeploy all applications in the monorepo since it was treated as a shared input. Now, Vercel inspects the lockfile’s contents to determine which applications have dependency changes, further reducing potential queue times. Learn more about skipping unaffected projects in monorepos. Read more

Groq, fal, and DeepInfra join the Vercel Marketplace

March 18, 2025

The Vercel Marketplace now has an AI category for tools to integrate AI models and services directly into Vercel projects. Groq, fal, and DeepInfra are available as first-party integrations, allowing users to: Seamlessly connect and experiment with various AI models to power generative applications, embeddings, and more Deploy and run inference with high-performance AI models, optimized for speed and efficiency Leverage single sign-on and integrated billing through Vercel, including new prepaid options for better cost control With prepaid plan options, users can now manage AI costs more predictably by purchasing credits upfront from a model provider. These credits can be used across any model offered by that provider. Explore the new AI category, read the docs, and get started with Groq, fal, and DeepInfra on the Vercel Marketplace, available to users on all plans. You can also explore the most popular models from each provider in the AI SDK playground. Read more

Vercel Firewall protects against the SAMLStorm vulnerability

March 18, 2025

We have deployed a proactive security update to the Vercel Firewall, protecting against a recently disclosed vulnerability in the xml-crypto package, dubbed SAMLStorm (CVE-2025-29774 and CVE-2025-29775). This vulnerability, which affects various SAML implementations, could allow attackers to bypass authentication mechanisms. What This Means for Vercel Customers Automatic protection with the Vercel Firewall: Vercel Firewall automatically mitigates this risk for you, but updating xml-crypto is still recommended Update xml-crypto: If you're using xml-crypto package 6.0.0 and earlier, or a package that depends on xml-crypto, update to 6.0.1, 3.2.1, or 2.1.6 for the patched versions We'll continue to monitor for new developments and provide updates as necessary See the SAMLStorm report for more details on the vulnerability, and reach out to Vercel Support if you have questions. Read more

Reduced Log Drains costs with smaller billable increments

March 17, 2025

We’ve updated Log Drains pricing on all Pro and Enterprise plans, reducing the charge increments. Data transferred for Log Drains will be billed at $0.50 per 1GB, instead of the previous news: [0 per 20GB, providing more precise usage tracking and better cost efficiency. Learn more about Log Drains. Read more

Vercel Marketplace integrations now available in v0

March 14, 2025

Users of v0—our collaborative AI assistant used to design, iterate, and scale full-stack applications—can now leverage integrations from the Vercel Marketplace, starting with Upstash, Neon, and Supabase. Install directly from the project sidebar or within v0’s chat interface. When added, these integrations redirect you to the Vercel Marketplace where you can configure environment variables, available to both Vercel and v0. Explore a example generation. Read more

Faster domain aliasing for large-scale multi-tenant applications

March 14, 2025

Bulk aliasing for multi-tenant applications now runs significantly faster, reducing total aliasing time by up to 95%. Multi-tenant applications on Vercel let a single project serve many customers behind the scenes. These applications are often fronted by hundreds or thousands of domains. Previously, aliasing—the process of pointing a domain to a different deployment—was a slow process that added significant overhead to deployments. This optimization is now live for all customers and has led to dramatic improvements, like: App with 13,254 domains: ~10min → 28 seconds App with 23,743 domains: 8min 37secs → 26 seconds Learn more about multi-tenant applications on Vercel. Read more

Jeanne DeWitt Grosser joins Vercel as COO

March 13, 2025

When I started Vercel, my vision was simple: make building for the web more accessible and more powerful. That belief has fueled Vercel’s growth, empowering developers to bring their biggest ideas to life. Today, we’re welcoming Jeanne DeWitt Grosser, former Chief Business Officer at Stripe, as Vercel’s Chief Operating Officer to help further this mission. Vercel is building the foundation to power the next billion developers. Achieving this vision requires strong leadership and operational excellence. As COO, Jeanne will lead our go-to-market function. Read more

Increased Hobby usage limits for Image Optimization

March 7, 2025

We've increased Image Optimization included usage for Hobby teams: Image Transformations: from 3K to 5K per month Image Cache Reads: from 180K to 300K per month Image Cache Writes: from 60K to 100K per month Learn more about Image Optimization pricing and its recent recent price reduction. Read more

Personalization strategies that power ecommerce growth

March 7, 2025

Personalization works best when it’s intentional. Rushing into it without the right approach can lead to higher costs, slower performance, and poor user experience. The key is to implement incrementally, with the right tools, while maintaining performance. When personalization is implemented effectively, it drives real business results, returning $20 for every news: [ spent and driving 40% more revenue. Let's look at what personalization is, how to implement it correctly, and why Next.js and Vercel achieve optimal outcomes. Read more

Overview page in Observability

March 6, 2025

Vercel Observability now includes an overview page that provides a high-level view of your application's performance. This new dashboard aggregates key metrics from Edge Requests, Fast Data Transfer, and Vercel Functions, giving you instant insights into request and data transfer volumes, as well as function performance. Each metric also serves as a starting point for deeper analysis, with one-click access to their dedicated dashboards for more detailed insights. Try it in your Observability dashboard. Read more

Vercel Firewall rule builder now supports `OR` for rule condition groups

March 5, 2025

The Vercel Firewall now supports using an OR operator to link condition groups within a custom WAF rule. Previously, customers could only use an AND operator to join condition groups. This update now supports AND or OR , allowing customers to create more complex WAF actions. Learn more about the Vercel Firewall or navigate to your Firewall tab to customize rules. Read more

How Fluid compute works on Vercel

March 3, 2025

Fluid compute is Vercel’s next-generation compute model designed to handle modern workloads with real-time scaling, cost efficiency, and minimal overhead. Traditional serverless architectures optimize for fast execution, but struggle with requests that spend significant time waiting on external models or APIs, leading to wasted compute. To address these inefficiencies, Fluid compute dynamically adjusts to traffic demands, reusing existing resources before provisioning new ones. At the center of Fluid is Vercel Functions router, which orchestrates function execution to minimize cold starts, maximize concurrency, and optimize resource usage. It dynamically routes invocations to pre-warmed or active instances, ensuring low-latency execution. By efficiently managing compute allocation, the router prevents unnecessary cold starts and scales capacity only when needed. Let's look at how it intelligently manages function execution. Read more

Using the AI SDK to build Sitecore Stream's AI-powered brand aware assistant

March 3, 2025

Sitecore—a leading digital experience platform—wanted to create a transformative AI tool that would help marketers connect more deeply with their brand assets, driving both consistency and creativity. Using the AI SDK, they lunched Sitecore Stream—a dynamic, AI-powered brand assistant that empowers marketers to interact with their brand content in visually interactive and conversational way. Read more

Improvements to Vercel Firewall system bypass rules

February 28, 2025

System bypass rules allow Pro and Enterprise customers to configure firewall rules to skip Vercel system mitigations, including DDoS protection, for specific IPs and CIDR ranges. Although we strongly recommend against disabling protections, customers—particularly ones that deploy a proxy in front of Vercel—may experience traffic issues that can be mitigated by deploying system bypass rules. Improvements to the system bypass rules give customers additional control over how the rules are deployed, including: Expanded support beyond production domains to preview domains Added support for single domain rules for preview deployment URLs and aliases Expanded project-scoped bypass rules to include all domains connected to a project Increased limits for system bypass rules for Pro to 25 and Enterprise to 100 (from 3 and 5 respectively) Learn more about the Vercel Firewall. Read more

Automatic pnpm v10 support

February 28, 2025

Vercel now supports pnpm v10. New projects with a pnpm-lock.yaml file with lockfileVersion: '9.0' will automatically use pnpm v10 for Install and Build Commands. Existing projects will continue to use pnpm v9 for backwards compatibility, since pnpm v9 also uses lockfileVersion: '9.0'. Check your build logs to see which version a deployment uses. If you'd like to manually upgrade or downgrade your version, use Corepack. Visit the package managers documentation to learn more. Read more

Fast Data Transfer for rewrites between your team's projects is now free

February 27, 2025

External rewrites between projects within the same team now use Fast Data Transfer only for the destination request. This change makes Fast Data Transfer for the original request free. Commonly used as a reverse proxy or for microfrontend architectures, rewrites can be configured in vercel.json, middleware, or next.config.ts to route requests between the same or separate Vercel projects without changing the URL shown to the user. Usage for external rewrites to the same team: Fast Data Transfer for the original and destination request have been optimized and consolidated into a single stream, reducing overall transfer. Each external rewrite triggers a full request lifecycle, including routing and Web Application Firewall checks, ensuring security policies are enforced per project, and counts as a separate Edge Request. Learn about rewrites and monitor your Fast Data Transfer usage and observability. Read more

Statsig joins the Vercel Marketplace

February 27, 2025

The Vercel Marketplace now has an Experimentation category to allow developers to work with feature flagging and experimentation providers in Vercel projects. Statsig—a modern feature management, experimentation, and analytics platform—is now available as a first-party integration in this new category, so users can: Connect Statsig with your Vercel projects directly from the Vercel Marketplace Leverage integrated billing through Vercel Sync your Statsig experiments into Edge Config for ultra-low latency Manage and roll out features progressively, run A/B tests, and track real-time results Additionally, you can use the Flags SDK to load experiments and flags from Statsig using the newly released @flags-sdk/statsig provider. Explore the template or get started with Statsig on the Vercel Marketplace, available to users on all plans. Read more

IP address details added in the Vercel Firewall dashboard

February 27, 2025

The Vercel Firewall dashboard now displays enriched IP address data, including the autonomous system (AS) name, AS number (ASN), and geolocation on hover. This information helps identify the origin of an attack, determine the owner of an IP address, and create targeted custom rules to block malicious traffic. Learn more about the Vercel Firewall. Read more

Middleware now supports Node.js

February 26, 2025

Middleware support for the Node.js runtime is now available, providing full Node.js support for authentication, personalization, and more—using familiar APIs. Middleware continues to be deployed globally on Vercel, regardless of the runtime used. We are first releasing support for Node.js Middleware in Next.js 15.2. This experimental feature requires the Next.js canary channel. Upgrade to next@canary and enable the nodejs experimental flag in your config to use it: You must also specify the Node.js runtime in your middleware file: Deploy now with Next.js 15.2. Read more

One-click linking from Usage to Vercel Observability dashboards

February 25, 2025

Metrics on the Usage dashboard now offer one-click access to corresponding Vercel Observability dashboards, making it easier to dive deeper into team and project usage. This new linking is available today for: Vercel Functions Edge Network Image Optimization Incremental Static Regeneration Builds (when viewing per project) Try it from your Usage dashboard and learn more about Vercel Observability. Read more

Granular branch matching for Git configuration in vercel.json

February 25, 2025

Vercel now supports glob patterns (like testing-*) in the git.deploymentEnabled field, giving you more control over branch deployments. Previously, you could disable deployments for specific branches by explicitly naming them. Now, you can use patterns to match multiple branches at once. For example, the configuration below prevents deployments on Vercel if the branch begins with internal-. Learn more about Git configuration. Read more

Changes to supported TLD registrations

February 25, 2025

We’ve updated our list of supported Top-Level Domains (TLDs) registrations, adding new options and removing select ones as we refine our domain offerings. Newly supported TLD registrations We now support 66 additional TLDs, including: - Generic domains (e.g. .page, .food, and .hosting) - Professional domains like (e.g. .lawyer, .phd, and .inc) - Lifestyle domains (e.g. .beauty, .living, and .lifestyle) - Interest-based domains (e.g. .guitars, .yachts, and .watches) TLD registrations no longer supported We have removed registration support for select TLDs, including: - Various country-code TLDs (ccTLDs, e.g. .at, .lu, .ma) - Regional TLDs (e.g. .berlin, .wales, .istanbul) - Multiple compound TLDs (e.g. .com.co, .org.pl, .co.nz) Future plans for TLD registration support We’re continuing to improve our domain offerings by: Enhancing support for country-code TLDs (ccTLDs), with plans to reintroduce select options. Expanding our portfolio with additional generic TLDs (gTLDs). These changes take effect immediately. Existing registrations, renewals, and services for deprecated TLDs remain unaffected. Read more

Integrating Vercel and Sitecore for 2x faster development times and 111% higher conversions

February 24, 2025

Avanade is the world’s leading expert on Microsoft which delivers AI-driven solutions for cloud, data analytics, cybersecurity, and ERP. The team at Avanade started on a comprehensive transformation, ultimately adopting Next.js, Vercel, and Sitecore XM Cloud to establish a modern, composable system capable of delivering highly responsive experiences to global clients. This migration helped replace monolithic systems with tightly coupled components and manual deployments, which hurt performance and feature rollouts. Read more

New monorepo projects now skip builds with unchanged code by default

February 24, 2025

Previously, we added opt-in support for skipping builds with unchanged code in monorepos to reduce build queueing. This behavior is now the default for new projects. To enable deployment skipping in an existing project, visit the Build and Deployment settings for the project. Additionally, this setting has been added to the Vercel provider for Terraform in 2.10.0. Learn more about skipping deployments. Read more

Observability for Edge Requests now includes more traffic parameters

February 24, 2025

We’ve expanded the Edge Request dashboard in Vercel Observability to show additional request data by: User agent IP address JA4 Referrer Hostname Available on all plans, these insights help you monitor traffic patterns and identify potential threats, which you can address using Vercel Firewall. Route-level data is available to Observability Plus customers. View your Edge Request dashboard and learn more about Vercel Observability. Read more

Consolidated Build and Deployment settings

February 21, 2025

We’ve simplified the Project Settings page, bringing all build customization options under a unified Builds and Deployment section. Vercel framework-defined infrastructure automatically detects settings for many frontend frameworks, but you can still customize build options to fit your needs. Learn more about project settings and how to configure a build. Read more

Vercel security roundup: Faster defenses and better visibility for your apps

February 21, 2025

Every second, Vercel blocks attacks before they reach your applications—keeping businesses online and developers focused on shipping, not security incidents. Vercel’s security capabilities combine real-time DDoS mitigation, a powerful Web Application Firewall (WAF), and seamless SIEM integrations to provide always-on protection without added complexity. Here’s what happened in the last quarter. Read more

Sync projects with @vercel/related-projects

February 20, 2025

The new @vercel/related-projects package helps sync deployment information across separate Vercel projects, ensuring your applications always reference the latest preview or production deployment URLs without manual updates or environment variable changes. Previously, developers had to manually enter deployment URLs, manage connection strings, or use environment variables to keep the projects communicating effectively. Now, this data is automatically available and updated at both build and runtime. For example, a monorepo containing: A frontend Next.js project that fetches data from an API An backend Express.js API project that serves the data Related Projects can now ensure that each preview deployment of the frontend automatically references the corresponding preview deployment of the backend, avoiding the need for hardcoded values when testing changes that span both projects. Related Projects are linked using a Vercel project ID. You can find your project ID in the project Settings page in the Vercel dashboard. Learn more about linking related projects. Read more

npm i flags

February 20, 2025

The Flags SDK—our open source library for using feature flags in Next.js and SvelteKit applications—is now available under the new package name flags. The new name signals our commitment to open source and the independence of the package from any specific entity or platform. Our framework-first approach of the SDK aims to simplify usage, avoid client-side flag evaluation, and improve user experience by eliminating layout shifts. We are working on adapters with partners like Statsig, Optimizely, and LaunchDarkly to ensure a seamless integration with the Flags SDK. Until now, each provider established their own approach to using feature flags in frameworks like Next.js, which led to duplicate efforts across the industry and drift in implementations. Going forward, the Flags SDK will help all feature flag and experimentation providers benefit from its tight integration to frameworks, while retaining their unique capabilities. If you are using @vercel/flags, make sure you are updating to version 3.1.1 and switch your imports and package.json to flags. Learn more in our redesigned documentation and examples. Read more

Improved traffic visibility on Firewall overview page

February 20, 2025

The Vercel Firewall overview page now shows improved visibility into your traffic and the Firewall status. Navigate to your Firewall page to see: Status of the system firewall A warning banner if a reverse proxy is inhibiting Vercel's ability to protect your site Tabbed view for easier traffic filtering Rules displayed below the chart with better readability The Vercel Firewall automatically mitigates DDoS attacks for all Vercel deployments. You can further secure your site with custom rules and IP blocking, and by turning on Attack Challenge Mode when under high-volume attacks. Learn more about the Vercel Firewall. Read more

New Observability dashboard for Image Optimization

February 19, 2025

Vercel Observability now includes a dedicated dashboard for Image Optimization, providing deeper insights into image transformations and efficiency. This update follows the introduction of a new pricing model, and includes: Transformation insights: View formats, quality settings, and width adjustments. Optimization analysis: Identify high-frequency transformations to help inform caching strategies. Bandwidth savings: Compare transformed images against their original sources to measure bandwidth reduction and efficiency. Image-specific views: See all referrers and unique variants of an optimized image in one place. This dashboard is available to customers on all plans and is compatible with both the new and legacy pricing models. View your Image Optimization dashboard and learn more about new pricing changes and Image Optimization. Read more

Deployment integration actions for Marketplace integrations

February 19, 2025

Marketplace integration providers can now register integration actions for deployments, allowing for automated resource-side tasks such as database branching, environment variable overrides, and readiness checks. When a user deploys a project that has connected Marketplace integration with configured actions, the deployment will pause and wait for all integration actions to complete successfully. This ensures that the deployed resources are properly set up before the deployment proceeds. Users will also receive helpful suggestions within the integration about which actions are available and should be executed. Learn more about integration actions. Read more

Observability for Vercel Functions now offers a quick-view of key insights

February 18, 2025

Observability's Vercel Functions dashboard now shows quick-view tiles with key metrics, such as: Active compute model, like Fluid compute, which enhances efficiency, minimizes cold starts, and optimizes performance Compute saved with Fluid compute enabled Average memory usage for your functions P75 Time to First Byte (TTFB) for performance monitoring Cold start frequency to track optimization impact These insights are available for all plans. Learn more about Observability and Fluid compute. Read more

Faster transformations and reduced pricing for Image Optimization

February 18, 2025

We’ve optimized our Image Optimization infrastructure, including: 60% faster transformations New, opt-in reduced pricing Previously, usage was measured by the number of unique source images ($5 per 1K source images). You can now opt into usage based on transformations with regional pricing, starting from: Image Transformations: $0.05 per 1K image transformations Image Cache Reads: $0.40 per 1M cache read units Image Cache Writes: $4.00 per 1M cache write units This new pricing model is opt-in through your project settings. There are no changes to existing customers using Image Optimization New projects for existing customers will also have no changes New customers will start on the new pricing today Pro and self-serve Enterprise customers can view the projected cost difference when enabling in settings. All Enterprise customers can also reach out to their account team to discuss new pricing. Hobby customers have been moved to the new model's included allotments. Learn more about Image Optimization pricing. Read more

Additional options for sharing deployments externally

February 14, 2025

You can now share deployments with external collaborators. Previously, invitations, access requests, and shareable links were limited to the preview URL for a branch or custom aliases. The share modal—accessible by selecting Share on a deployment page or from the Vercel Toolbar menu—now allows sharing the specific deployment you are on or the always up-to-date preview URL for the branch. Read more about sharing deployments. Read more

Automated DNS configuration with Domain Connect

February 14, 2025

Vercel now supports Domain Connect, an open standard that simplifies DNS configuration. With one click, you can set up your domain without manually copying DNS records—saving time and reducing errors. Cloudflare-managed domains are now supported with more providers coming soon. To get started: Add a new domain to your Vercel project, and Vercel will detect if your domain qualifies for setup through Domain Connect, prompting you to proceed automatically or configure it manually. We're also implementing Domain Connect as a DNS provider, enabling external services to configure Vercel Domains just as easily. Learn more about Vercel domains. Read more

Support for React Router v7

February 13, 2025

Vercel now supports React Router v7 applications when used as a framework: This includes support for server-rendered React Router applications using Vercel's Fluid compute. Further, the Vercel preset intelligently splits application bundles across Vercel Functions, and supports custom server entry points. Deploy React Router to Vercel or learn more about React Router on Vercel. Read more

Redeploy without leaving project settings

February 12, 2025

When updating project settings, such as environment variables, Vercel will now automatically prompt you to redeploy. A toast notification will appear when you change any settings that require a redeploy to take effect. After clicking Redeploy, you can track the progress of your deployment. Learn more about project settings. Read more

Bridging the gap between design and code with v0

February 12, 2025

Speakeasy specializes in building developer-focused SDKs—to help developers build their products. They adopted v0 to bridge the workflow from design to code, using it to accelerate rapid prototyping and reduce implementation time. Read more

Manage multiple Vercel Function regions in the dashboard

February 12, 2025

Pro and Enterprise plans can now select multiple regions for Vercel Functions directly from the dashboard. This update simplifies configuration by removing the need to define regions in vercel.json. Multi-region support is available for all Vercel Functions and supports Vercel's implementation of Fluid compute, which encourages a dense global compute model that positions dynamic functions closer to your data. Visit your project’s Settings tab to customize your regions or learn more about configuring regions for Vercel Functions. Read more

Split-tgz is now the default CLI archive deployment behavior

February 11, 2025

Archive deployments are useful for deploying large projects with thousands of files from the CLI. We previously released the split-tgz archive deployment as a new archive option: vercel deploy --archive=split-tgz. This new capability offered up to 30% faster uploads and avoided file upload size limits. We’ve confirmed split-tgz’s stability and made it the default behavior for tgz. This means the separate split-tgz option is now deprecated as the split-tgz functionality and benefits power the default tgz option. Learn more about CLI archive deployments. Read more

Vercel database templates now support any marketplace provider

February 11, 2025

We’ve updated our database starter templates to support selecting any Postgres or Redis provider available in the Vercel Marketplace when deploying. These templates are now provider-agnostic, allowing developers to seamlessly integrate alternative database and key-value store solutions while maintaining the same developer experience. Redis: Nuxt.js, Next.js, SvelteKit Postgres: Next.js, Nuxt.js, SvelteKit, Kysely, Drizzle Check out the documentation to learn how to deploy your own. Read more

Enhanced firewall data now available in Monitoring

February 7, 2025

Monitoring now has better firewall support, offering insights into your firewall rules: Filter blocked requests by actions and custom firewall rules More fields are now displayed when available: IP Country User Agent Route Request Path Region These metrics are available for all Observability Plus and Monitoring customers. Monitoring recently became part of Observability Plus. Read more

Faster deploy times for large builds

February 6, 2025

We optimized the deploy step of the build process to reduce build times by 2.8 seconds at P99, 760ms at P75, and 410ms on average. For customers with a large number of Vercel Functions (100+), builds are more than 50 seconds faster. Several customers have time savings of over 2 minutes. Check out the documentation to learn more about builds. Read more

New execution duration limit for Edge Functions

February 6, 2025

Note: This changelog entry is historical. Edge Functions are deprecated for new projects. Use Vercel Functions with the Node.js runtime and Fluid compute. Use Routing Middleware for request-time routing before a response completes. Starting on March 1st, 2025, we will begin the rollout of a new execution duration limit of 300 seconds for Vercel Functions using the Edge runtime. Previously, functions using the Edge Runtime had no fixed timeout for streaming responses, leading to unpredictable behavior based on system resources and traffic. With this update, functions using the Edge Runtime will consistently allow streaming responses for up to 300 seconds, including post-response tasks like waitUntil(). Learn more about Vercel Functions using the Edge runtime. Read more

Dark mode, expanded search, and more repositories in Grep

February 5, 2025

We've made improvements to Grep, our tool for quick code search. You can now search across 1,000,000 public git repositories The app has been rebuilt with Next.js 15, improving performance with Partial Prerendering Support for dark mode Try Grep today. Read more

Deployment pages now display key configuration settings

February 5, 2025

Project Overview and Deployment Details pages now include a Deployment Configuration section under the deployment card. Expand to view snapshots of Fluid Compute, Function CPU, Deployment Protection, Skew Protection, and Secure Compute settings. This section is available for all new deployments moving forward. It will appear on your Project Overview page after your next production deployment. Read more

Introducing Fluid compute

February 4, 2025

While dedicated servers provide efficiency and always-on availability, they often lead to over-provisioning, scaling challenges, and operational overhead. Serverless computing improves this with auto-scaling and pay-as-you-go pricing, but can suffer from cold starts and inefficient use of idle time. It’s time for a new, balanced approach. Fluid compute evolves beyond serverless, trading single-invocation functions for high-performance mini-servers. This model has helped thousands of early adopters maximize resource efficiency, minimize cold starts, and reduce compute costs by up to 85%. Read more

Vercel Functions can now run on Fluid compute

February 4, 2025

Vercel Functions can now run on Fluid compute, bringing improvements in efficiency, scalability, and cost effectiveness. Fluid is now available for all plans. What’s New Optimized concurrency: Functions can handle multiple requests per instance, reducing idle time and lowering compute costs by up to 85% for high-concurrency workloads Cold start protection: Fewer cold starts with smarter scaling and pre-warmed instances Optimized scaling: Functions scale before instances, moving beyond the traditional 1:1 invocation-to-instance model Extended function lifecycle: Use waitUntil to run background tasks after responding to the client Runaway cost protection: Detects and stops infinite loops and excessive invocations Multi-region execution: Requests are routed to the nearest of your selected compute region for better performance Node.js and Python support: No restrictions on native modules or standard libraries Enable Fluid today or learn more in our blog and documentation. Read more

Enterprise teams can now ship faster without build queues

January 31, 2025

On-demand concurrent builds automatically and dynamically scale builds, increasing build capacity and shipping velocity. Starting today, new projects in Enterprise teams will use on-demand concurrency by default to eliminate build queue bottlenecks. You can turn this feature on for existing projects at any time with urgent on-demand concurrent builds or enable it at the project level. You are charged for on-demand concurrency based on the number of 10-minute build slots required to allow the builds to proceed as explained in usage and limits. Check out the documentation to learn more about on-demand concurrent builds . Read more

ISR on Vercel is now faster and more cost-efficient

January 30, 2025

When Next.js introduced Incremental Static Regeneration (ISR) in 2020, it changed how developers build for the web. ISR combines the speed of static generation with the flexibility of dynamic rendering, enabling sites to update content without requiring full rebuilds. Vercel has supported ISR from day one, making it easy for teams at The Washington Post, Algolia, and Sonos to serve fresh content while keeping page loads fast. Read more

Incremental Static Regeneration (ISR) is now faster and more cost-efficient

January 30, 2025

Incremental Static Regeneration (ISR) enables you to update content in the background without needing to redeploying your application. You can scale CMS or content-backed applications to millions of pages without having slow builds. We've optimized our infrastructure to make ISR faster and more cost-efficient: Smaller writes: ISR cache writes are now compressed by default, using fewer ISR write and read units (8KB chunks) per update and lowering Fast Origin Transfer (FOT) costs. Both reads and writes are now compressed. Region-aware caching: The ISR cache is now available in all regions and automatically aligns with your functions' region. If your project spans multiple regions, the most cost-effective location is chosen automatically. This improves performance, especially for traffic outside North America, and regional pricing applies. Redeploy your project to apply these updates or learn more about ISR. Update: The rollout of this change completed on February 5th, 2025 around 8am PST. Read more

Edge Function metrics now available in Monitoring

January 30, 2025

Note: This changelog entry is historical. Edge Functions are deprecated for new projects. Use Vercel Functions with the Node.js runtime and Fluid compute. Use Routing Middleware for request-time routing before a response completes. Monitoring now includes three new metrics for Edge Functions to provide a comprehensive view of your Edge Function activity and performance: Edge Function Invocations**: **Tracks the total number of times your Edge Functions are invoked, including both successful and errored calls Edge Function Execution Units**: **Measures the CPU time your Edge Functions use, calculated in 50ms increments Fast Origin Transfer (Incoming and Outgoing)**: **Track data transfer rates to and from your origin servers These metrics are available for all Observability Plus and Monitoring customers. Monitoring recently became part of Observability Plus. Read more

Filter for your own requests in Logs

January 29, 2025

You can now filter logs to display only requests made from your browser. This simplifies debugging by isolating your requests in high-traffic environments. It matches your IP address and User Agent to incoming requests. Visit your project's Logs tab and toggle the user filter to get started or learn more about runtime logs. Read more

Clients blocked by persistent actions now receive a 403 Forbidden response

January 28, 2025

Starting today, when the Vercel Web Application Firewall (WAF) blocks a client with a persistent action, it will respond with a 403 Forbidden status instead of failing silently. This change now makes it clear that the connection is being intentionally denied. Persistent actions in the WAF help reduce edge request load and stop malicious traffic earlier, cutting down unnecessary processing for your applications. Learn more about persistent actions. Read more

Working with Figma and custom design systems in v0

January 27, 2025

v0’s ability to import existing Figma files allows designers and developers to bridge the gap between design tools and AI-driven development. This feature extracts context from Figma files, along with any supplementary visuals, and passes them into v0's generation process. Read more

Mitigating Denial of Wallet risks with Vercel

January 24, 2025

Unlike traditional cyberattacks that target code or infrastructure vulnerabilities, Denial of Wallet (DoW) attacks focus on draining a service's operational budget. At Vercel, we're building controls and anomaly detection to help you defend against these threats and protect your applications. Read more

Project settings are now searchable

January 24, 2025

You can now search within project settings in the Vercel Dashboard, making it easier to quickly find a specific setting. To get to your project settings: Select a project from your Team Overview page Select the Settings tab. Learn more about project settings. Read more

Firefox extension for Vercel Toolbar

January 24, 2025

The Vercel Toolbar extension is now available for Firefox, in addition to Chrome. With this extension you can use the Vercel Toolbar on your production deployments, set preferences for when the toolbar appears and activates, and drag and release to add a screenshot of a selected area to a comment. Install the Firefox extension from the Firefox Browser Add Ons page to get started or visit our documentation to learn more about the Vercel Toolbar and browser extensions. Read more

Preview your site's Firewall status and Web Analytics from the Project Overview

January 24, 2025

The project overview page now shows a preview of your production traffic and firewall status. The Vercel Firewall automatically mitigates DDoS attacks for all Vercel deployments. You can further secure your site with custom rules and IP blocking, and by turning on Attack Challenge Mode when under high-volume attacks. On the project overview page you'll see the status of the firewall, requests blocked and challenged in the past 24 hours, and a warning if a reverse proxy is inhibiting Vercel's ability to protect your site. Vercel Web Analytics gives you insight into your site's visitors and traffic. When the feature is enabled, you'll see your site's traffic on the project overview page. Learn more about Vercel Web Analytics and Vercel Firewall. Read more

Vercel acquires Tremor to invest in open source React components

January 22, 2025

Tremor is an open source library built on top of React, Tailwind CSS, and Radix. It consists of 35 unique components and 300 blocks that can be copy-pasted to build visually rich and interactive dashboards. The Tremor community has seen impressive growth with over 16,000 stars, 300,000 monthly downloads, and 5,500,000 installs to date. Today, Tremor and its cofounders Severin Landolt and Christopher Kindl are joining Vercel’s Design Engineering team where they'll be working on UI components for the Vercel Dashboard, v0, and more. This acquisition strengthens our commitment to open source and providing developers with the best tools for building exceptional user interfaces. Read more

Self-serve domain renewals and redemptions now available

January 21, 2025

Self-serve domain renewals and redemptions are now available in the Vercel dashboard. Previously limited to automatic renewals to ensure uninterrupted service, domains can now be manually renewed at your convenience with just a few clicks, directly from the dashboard. Domain renewals To renew your domains directly: Navigate to your team's Domains tab Click the three dots next to the domain you want to renew Select Renew Additionally, you may click the Renew Domain button on any domain detail page. Domain redemptions For expired domains with a redemption period (typically 30 days), you can now recover them directly in the dashboard: Start the redemption process on the domain detail page A redemption fee will be applied, depending on the domain registry Read our domains renewal documentation for more information. Read more

AI SDK 4.1

January 20, 2025

The AI SDK is an open-source toolkit for building AI applications with JavaScript and TypeScript. Its unified provider API allows you to use any language model and enables powerful UI integrations into leading web frameworks such as Next.js and Svelte. Read more

Claim Deployments now available for fast and secure deployment transfers

January 20, 2025

Multi-tenant platforms, like AI agents and visual building apps, can now easily transfer deployment ownership directly to users or teams. How it works: Deployment creation: Any third-party can create a new deployment using the Vercel CLI or using the Vercel API: POST /files and POST /deployments Initiate transfer: The Vercel API endpoint is then used to generate a claim-deployment URL for that deployment. User confirms their team: The user selects their Vercel team and completes the transfer. Check out our documentation to learn more. Read more

Node.js 18 is being deprecated on September 1, 2025

January 16, 2025

Following the Node.js 18 end of life on April 30, 2025, we are deprecating Node.js 18 for Builds and Functions on September 1, 2025. Will my existing deployments be affected? No, existing deployments with Serverless Functions will not be affected. When will I no longer be able to use Node.js 18? On September 1, 2025, Node.js 18 will be disabled in project settings. Existing projects using 18 as the version for Functions will display an error when a new deployment is created. How can I upgrade my Node.js version? You can configure your Node.js version in project settings or through the engines field in package.json. How can I see which of my projects are affected? You can see which of your projects are affected by this deprecation with: Read more

CLI archive deployments are now up to 30% faster with split-tgz archive option

January 16, 2025

The archive option was introduced for CLI deployments hitting rate limits like the limit on the maximum amount of files. Prebuilt deployments commonly use archive uploads as they generate thousands of files at build time. Previously, archive deployments were always compressed into one large file with the only existing --archive option, tgz. Deployments using tgz may hit the file size upload limit. Additionally, uploading one large archive file is slower than uploading multiple file parts. The beta split-tgz format resolves these issues by splitting large archives into smaller parts. split-tgz avoids the static file upload limit and uploads large prebuilt projects up to 30% faster. Example usage: vercel deploy --archive=split-tgz Learn more about CLI deployments. Read more

Audit logs with SIEM integration now generally available

January 16, 2025

Audit logs are now generally available for Enterprise customers, and can be integrated with SIEMs for real-time export. Audit logs provide an auditable trail of key events and changes within your Vercel team. With an immutable record, you can track who performed an action, what was done, and when—with access to up to 90 days of historical data. Enterprise customers can also configure a real-time audit log stream to their existing Security Information and Event Management (SIEM) tools, such as Datadog or Splunk. Additionally, logs can be sent to durable object storage solutions like Amazon S3, Google Cloud Storage, or a custom HTTP POST endpoint. For more details, check out the Audit Log documentation or contact your account manager. Read more