liuzhao1225 /YouDub-webui

About YouDub-webui

Point YouDub WebUI at a YouTube or Bilibili link, or upload a local video, and it runs the full translation and dubbing pipeline on your machine. It transcribes, translates, clones a voice and mixes the dub with the original music. The project is an MVP, and translation needs a paid OpenAI-compatible API key.

Best for: Small dubbing teams who want to run video translation on their own hardware.

• Takes a URL or a local file and returns a finished mp4
• Picks one of three outputs: hard subs, a dub, or both
• Splits vocals from music with Demucs so dubs keep the original sound
• Transcribes with Whisper, then translates through an OpenAI-compatible API
• Keeps a task history with stage status, logs, resume and rerun
• Stores SQLite data, cookies, logs and videos on your own machine

1. 1
   Install Python 3.12, Node.js 20+, and FFmpeg/ffprobe on your PATH. On Windows, use a shared FFmpeg build.
2. 2
   Clone the repo and pull the submodules: the command below, cd YouDub-webui, the next command

   git clone https://github.com/liuzhao1225/YouDub-webui.git

   Copy

   git submodule update --init --recursive

   Copy
3. 3
   Make a virtualenv and install deps: python3.12 -m venv .venv, the command below, (cd apps/web && npm ci). CUDA users install requirements-pytorch-cu128.txt first.

   .venv/bin/pip install -r requirements.txt

   Copy
4. 4
   Copy env.txt.example to .env, set FFMPEG_PATH and FFPROBE_PATH, add your Argon2id login hash, and fill in OPENAI_BASE_URL, OPENAI_API_KEY and OPENAI_MODEL.
5. 5
   Start the backend with the command below and the frontend with the next command

   .venv/bin/uvicorn backend.app.main:app --reload --host 0.0.0.0 --port 8000

   Copy

   npm --prefix apps/web run dev -- --hostname 0.0.0.0 --port 3000

   Copy
6. 6
   Open http://localhost:3000, log in, set cookies, proxy and API key in Settings, then submit a URL or upload a video.