pyVideoTrans
Desktop, web and CLI tool that transcribes, translates, dubs and re-muxes a video into another language.
Speech recognition, subtitle translation, AI voices and video merge in one free desktop app.
pyVideoTrans runs the whole chain of recognising speech, translating subtitles, synthesising a dub and merging it back into the video. It suits creators and localisers who want to pick their own models and keep the files on their machine.
What it is
You feed it a video and it runs speech recognition, subtitle translation, text-to-speech and video synthesis, with a pause at each stage so you can proofread. It can also work as separate tools for audio-to-SRT, subtitle translation and text-to-speech, and it offers speaker diarization plus voice cloning through models such as F5-TTS, CosyVoice and GPT-SoVITS.
Who it's for
It suits YouTubers, course makers and subtitle teams who want to mix local models with online APIs and edit each stage by hand. It is a poor fit if you want a polished hosted service with support, or if you cannot check the licences of the third-party models you plug in.
Key features
One-click video translation
Chains recognition, subtitle translation, speech synthesis and video merge into one run.
Pause and proofread stages
You can stop and hand-edit the transcript, translation and dub before continuing.
Multi-speaker dubbing
Speaker diarization lets you give each speaker a different AI voice.
Voice cloning models
Plugs into F5-TTS, CosyVoice and GPT-SoVITS for zero-shot voice cloning.
GUI, WebUI and CLI
Use the desktop app, a browser interface on port 7860, or a headless command line.
Vocal separation toolkit
Includes vocal separation, subtitle merging and audio-video alignment helpers.
Up and running in 3 steps
Our honest verdict
What's great
- Ships a pre-packaged Windows .exe, so Windows users skip the Python setup
- Lets you swap the recogniser, translator and voice engine independently, from local Faster-Whisper and Ollama to online APIs
- Offers stage-by-stage manual correction instead of a single black-box output
- Has a CLI and a Docker WebUI for batch or server use
Where it falls short
- Local recognition and voice cloning want an NVIDIA GPU with CUDA 12.8 and cuDNN 9.11 for decent speed; AMD acceleration goes through a separate Whisper.NET setup
- macOS and Linux users must install from source with Python 3.10, FFmpeg and uv
- Paid translation or TTS APIs are billed by their vendors, and each local model has its own licence you must check for commercial use
- The documentation site and community Q&A are largely Chinese-first
Pricing & license
Free under GPL-3.0; the project site states there is no subscription, sign-up or licence code. Your only costs are hardware and any paid translation or voice APIs you connect, which are billed by those vendors.
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Is pyVideoTrans free for commercial work?
The code is GPL-3.0 and the project site says there are no fees or sign-up. The README adds that integrated local models and online APIs carry their own terms, so check those before commercial use.
Do I need a GPU?
The site says it runs with or without an NVIDIA card. A card with CUDA speeds up local recognition, and the README names CUDA 12.8 and cuDNN 9.11 for that.
Can it run on a server?
Yes. The README documents a CLI for headless runs and a Gradio-based WebUI, with a Docker image exposing port 7860.
