# ElevenLabs opens Dubbing v2 to developers via API By Steven Van · 2026-08-06 The performance-preserving dubbing model from ElevenCreative is now callable through ElevenAPI, priced per minute with no charge for failed jobs. [ElevenLabs]() has released Dubbing v2 on ElevenAPI, its developer platform, announced on August 6, 2026. The model that impressed studios and localization teams when it launched in ElevenCreative earlier this year can now be called programmatically, which means any app, media pipeline, or creator platform can dub audio and video into more than 90 languages while keeping the original speaker's voice, emotion, and timing intact. [Source post](https://x.com/i/status/2085383576959197680) ## From studio feature to programmable dubbing When ElevenLabs introduced Dubbing v2 in late May 2026, it was only available inside ElevenCreative and through ElevenProductions, the company's managed localization service. API access was listed as coming soon, and developers who wanted it were pointed to a sales contact form. That gap is now closed. Dubbing v2 runs on ElevenAPI as the default model for automatic dubbing, so the same engine powering the web product is available to anyone with an API key. Teams that were stitching together transcription, translation, and text to speech calls to fake a dubbing pipeline can replace that whole chain with one workflow. ## Why v2 sounds different: it dubs the performance, not the transcript Traditional AI dubbing, including the first generation of ElevenLabs' own dubbing product, worked from a transcript. The audio was transcribed, translated, and re-synthesized, which produced flat delivery disconnected from how the original speaker actually sounded. Dubbing v2 conditions directly on the original performance. It analyzes emotion, intonation, pacing, pauses, and delivery style in the source audio and carries those characteristics into the translated output. Automatic voice cloning preserves each speaker's identity, pitch, and tone, and speaker separation detects multiple speakers even when their speech overlaps. A sync-aware translation system adapts the wording for spoken delivery and aligns starts, stops, and pacing with the original, cutting down the manual timing fixes that used to eat localization hours. ## What the API workflow looks like The developer flow is project based. You create a dubbing project with a source URL, a source language, and a target language, then poll the project as it moves from pending to ready to completed. Language targets can be added per project, so one upload can fan out into many dubbed versions. When generation finishes, the API returns a signed URL to the output, including a lossless audio option for production use. Official SDKs for Python and TypeScript wrap the endpoints, and ElevenLabs publishes a dubbing quickstart cookbook with working examples in both. Developers can also tune cloning strength, a setting that balances voice similarity against natural delivery in the target language, and the legacy v1 model remains available for anyone who needs to match older outputs. ## Pricing and what this unlocks for creator platforms Dubbing v2 is priced per minute of dubbed output on automatic dubbing, and failed jobs are not charged, which matters when you are processing user uploads at scale. For creator-facing products the use cases are immediate: video platforms can offer one-click multilingual versions of uploads, podcast tools can ship episodes in Spanish and Japanese from the same feed, course platforms can localize entire libraries, and social publishing tools can generate dubbed cuts for regional audiences automatically. Because the voice and emotional performance survive translation, the dubbed content still sounds like the creator rather than a generic synthetic narrator, which has been the main reason audiences bounce off AI dubs. ## The bigger picture: localization as an API call ElevenLabs has been steadily converting its studio products into developer primitives, and dubbing was the most requested holdout. With Dubbing v2 on ElevenAPI sitting alongside its text to speech, music, sound effects, and transcription endpoints, the company is positioning ElevenAPI as a full audio stack for products, not just a voice generator. For developers, the practical takeaway is simple: professional-grade dubbing that used to require translators, voice actors, and audio engineers is now a REST call, and the quality bar just moved. ![ElevenLabs]() ElevenLabs The most realistic AI audio platform. [View ElevenLabs →]() Sources [ElevenLabs on X](), [ElevenLabs blog: Introducing Dubbing v2](), [ElevenLabs Dubbing docs]() [Read ElevenLabs opens Dubbing v2 to developers via API on Creators Toolbox]() --- Canonical source: https://creatorstoolbox.com/blog/elevenlabs-dubbing-v2-api