Sarvam AI is what makes our Indian language support possible. Their Saaras v3 model transcribes audio across all 22 scheduled Indian languages with accuracy that Whisper simply can't match for Indic content. Bulbul v3 gives us 37+ natural Indian voice speakers for text-to-speech, and their translation API handles all 22 languages. We integrated their full stack STT, translation, TTS, transliteration, and language detection so creators can edit videos in Hindi, Tamil, Telugu, Kannada, Bengali, Malayalam, and 16 more languages. Free credits on signup, and the API just works.
Whisper powers our local transcription word-level timestamps, multiple model sizes (tiny to large-v3), all running on-device. It's the backbone of our edit-by-text feature: transcribe your video, then edit it like a document. No API calls, no per-minute billing.
SDXL Turbo lets users generate images from text prompts directly inside the video editor and drop them on the timeline. Running locally means unlimited generations with zero cost and full privacy.
XTTS v2 powers our voice cloning clone any voice from a 6-second sample, then generate voiceovers in that voice. Multilingual, open-source, and runs entirely local. This model is what makes our TTS feature possible.
Our in-browser AI features use Hugging Face Transformers.js. Running ML models directly in the browser without a backend call that's the magic Hugging Face enables for local-first AI apps.
We switched our monorepo to Bun for package management and dev server. Install times dropped dramatically, and bun dev with Turbopack gives near-instant hot reloads even on a codebase with 80+ dependencies. It just works.
Ollama powers our local LLM commands users can control the video editor in plain English like "remove the intro" or "speed up the middle." It runs entirely on the user's machine with no API keys needed. Ollama made local AI feel as easy as running a Docker container.
OpenCut AI's entire frontend runs on Next.js with Turbopack. It handles real-time canvas rendering, multi-track timeline state, and a full editor UI all with zero config. The dev experience with Turbopack hot reload made iterating on a complex video editor actually enjoyable.
Claude Code was the backbone of building OpenCut AI. We added 8+ AI features transcription, voice cloning, filler removal, AI image generation, natural language editing on top of an existing open-source video editor, and Claude Code's deep codebase understanding made it possible. Unlike Cursor or Copilot, it doesn't just autocomplete lines it reasons across the entire project, refactors confidently, and handles multi-file changes spanning a Next.js frontend and FastAPI AI backend in one shot. The 1M context window meant we could feed it our whole architecture and get coherent, production-ready code back. Nothing else came close for a project this ambitious
Free AI Video Editor OpenCutAI