How Factory scaled its cloud backend to one billion monthly requests on Vercel
By Steven Van ·
Factory on Vercel One billion backend API requests served daily 350ms p95 response time Scaled backend, internal tooling, and security without a dedicated…
Three years into building on Vercel, Factory now runs a single Next.js backend that handles one billion API requests a day across its entire platform, from individual developers to enterprise customers, at a p95 response time of 350ms or below, according to a Vercel case study.
- API routes, middleware, webhook handlers and log drain pipelines all run in the same Next.js app rather than separate infrastructure. Fluid compute keeps functions warm to avoid cold starts, and the only tuning Factory needed as traffic grew was to memory limits and function duration.
- Non-technical staff, such as account executives and solution architects, now build and deploy their own internal tools using Factory's Droids, which call Vercel's APIs directly to handle configuration, builds and deployments. That produces dozens of deployments a day, including internal analytics dashboards and customer-specific demos.
- After opening self-serve signups exposed the backend to bot signups and traffic from bad IPs, Factory added Vercel Observability to track traffic patterns and problem IPs, and uses the Web Application Firewall for route-level blocking alongside continuous rate limiting and DDoS protection.
Vercel
The platform for frontend developers — deploy, preview, and scale web apps and AI agents with zero config.
View Vercel →