pyVideoTrans 🧩 Workflow Open source

What it is

pyVideoTrans by jianchang512 (GPL-3.0) translates videos from one language to another. It transcribes speech with Whisper-family models, translates subtitles with an LLM or translation API, generates dubbed audio with a choice of TTS engines, and merges new subtitles and audio back into the video. It also works as a standalone transcriber, subtitle translator or batch TTS tool, with a desktop GUI and CLI. Updates land several times a month, and the project documents each supported engine with setup notes and limits.

Key features

• Speech recognition with faster-whisper, OpenAI Whisper and cloud ASR options
• Subtitle translation with ChatGPT, Claude, Gemini, DeepSeek, Ollama and other LLM or translation APIs
• Dubbing with Edge-TTS, Azure, F5-TTS, CosyVoice and other engines
• Voice cloning options to keep the original speaker's timbre
• Batch processing and separate subtitle/audio tools
• Windows package plus source install on macOS and Linux

How to install

On macOS the README installs prerequisites with Homebrew and uv, then git clone https://github.com/jianchang512/pyvideotrans.git and syncs dependencies with uv; Windows users can download a prebuilt package.

Pricing and rights

Free under GPL-3.0. Local models cost nothing; cloud ASR, translation and TTS services bill separately. Voice cloning should only be used with the speaker's consent.

Who it is for

YouTubers, educators and localisation freelancers who want to publish multilingual versions of their own videos without paying per-minute dubbing fees.

Verdict

A powerful, very active free alternative to paid dubbing tools. The interface is dense and documentation is mainly Chinese, and quality depends on which engines you pick.