jianchang512 /pyvideotrans

About pyvideotrans

pyVideoTrans takes a video in one language and puts it out in another. Speech recognition, subtitle translation, dubbing and assembly run as one chain. You can also use just the subtitle or text-to-speech parts. Some channels want paid API keys, and some local voices need a separate install.

Best for: Subtitling teams using local models or paid APIs to translate videos they already have.

• Runs speech recognition, subtitle translation, dubbing and video assembly in one go
• Turns batches of audio or video into SRT subtitles, with speaker separation
• Gives each speaker their own AI dubbing voice
• Clones a voice from a short sample with F5-TTS, CosyVoice or GPT-SoVITS
• Lets you pause and fix text at the recognition, translation and dubbing steps
• Ships a CLI for batch runs, a browser WebUI and a Docker image

1. 1
   On Windows, download the pre-packaged release, unzip it to a path with no Chinese characters or spaces, and run sp.exe.
2. 2
   For a source install, get Python 3.10 and FFmpeg (on Ubuntu/Debian: the command below).

   sudo apt-get install ffmpeg libsndfile1-dev

   Copy
3. 3
   Install uv, clone the repo, then run uv sync (add --all-extras for optional channels).
4. 4
   Start the desktop app with uv run sp.py.
5. 5
   Or run a CLI job: the command below.

   uv run cli.py --task vtv --name "./video.mp4" --source_language_code zh-cn --target_language_code en --voice_role "en-US-GuyNeural"

   Copy
6. 6
   For browser access, run uv sync --extra webui then uv run webui.py.