RepoPicks Radar

Updated every day around 23:30 UTC

Guides

Voice-Pro

abus-aikorea/voice-proReviewed by AI

Checked by RepoPicks Radar on

What is Voice-Pro?

A dubbing workbench that opens in your browser but runs on your machine. It downloads a YouTube video or takes your own file, separates the voice from the music, transcribes it with Whisper, translates the subtitles into one of 100+ languages and reads the result aloud with Edge-TTS, kokoro, F5-TTS, E2-TTS or CosyVoice, including cloning a voice from a short sample. You start it with a script that installs Python and the models by itself (about 10 GB), and you need an NVIDIA graphics card with 4 GB or more.

Voice-Pro is free and open source (GPL-3.0) and can replace ElevenLabs, Descript, and Rask AI.

How easy is it to install Voice-Pro?

D · terminal, several tools or a specific graphics card · Windows

  • Windows D
  • macOS D
  • Linux D

Windows is the only system the author says he has tested, and it is a D: you clone or unzip the source, run `start.bat`, and the first run downloads Python, the dependencies and about 10 GB of models, with an NVIDIA card of 4 GB or more and 20 GB of free disk space. Mac and Linux have the same `configure.sh` and `start.sh` scripts, so the same grade, but the README says their operation has not been verified.

Coming soon: install in 1 click with the RepoPicks app.

Voice-Pro on GitHub

Stars
13,421on Nov 10, 2026
Downloads
—all releases
Growth
+20stars in 6 days

Stars over time

13,40513,41013,41513,42013,425 Nov 5Nov 6Nov 7Nov 8Nov 9Nov 10 13,40513,41013,41513,42013,425 Nov 5Nov 8

Facts

Runs on
Windows, macOS, Linux
You need to install
Python, NVIDIA graphics card, FFmpeg
To build from source
Python, NVIDIA graphics card, FFmpeg
Disk space
1.2 GB–5.1 GB · estimate
Accounts or keys
Microsoft Azure Speech (optional, better voices), Microsoft Azure Translator (optional, instead of the free Google endpoint)
Cost
Free · README: "We have made all Voice-Pro code open source and completely free." Older versions had a 60-second limit with a paid subscription; that is gone. Optional Microsoft Azure translation and speech use your own paid Azure account.
License
GPL-3.0 (Copyleft)
Last change on GitHub
Jul 13, 2026

Voice-Pro risk signals (6)

  • Attentionships celebrity voice samples

    The README has tables of reference voices for cloning made of real public figures, with files such as `celebrities30sREADME/English/Donald Trump.jpg`, `Elon Musk.jpg`, `Joe Rogan.jpg`, `celebrities30sREADME/Korean/IU.jpg` and `celebrities30sREADME/Japanese/Ayase Haruka.jpg`, plus the feature line "Celeb voice podcasts". Making a real person say words they never said can break voice and likeness rights and, in several countries, impersonation laws.

    AI review

  • Attentiondownloads videos from youtube

    The first feature of the Dubbing Studio tab is "YouTube video downloads & audio extraction" using yt-dlp (`"yt-dlp>=2025.9.26"` in pyproject.toml). Downloading from YouTube goes against its terms of service, and re-dubbing someone else's video is a copyright matter in most countries.

    AI review

  • Attentionreads your browser cookies

    pyproject.toml lists `browser_cookie3` as a core dependency. That library reads the cookie database of browsers installed on the computer, which is how download tools reach videos that need you to be signed in; the README never mentions it.

    AI review

  • Infoauthor paused development

    README, twice: "Due to [WeConnect](https://www.wctokyoseoul.com) development work, Voice-Pro development and updates are not possible for the time being." The last commit and the last release (v4.0.0) in the snapshot are from 2026-07-13, 114 days before the check.

    AI review

  • Inforeadme contradicts itself on systems

    System Requirements says "OS: Windows 10/11 (64-bit), Linux, Mac (Apple Silicon)", while the note at the top says "It works well on Windows with NVIDIA GPU. Operation on Mac and Linux has not been verified." In pyproject.toml the GPU packages are only selected for Linux and Windows.

    AI review

  • Inforeadme carries marketing and old paid plans

    The README calls itself "The best AI speech recognition, translation, and multilingual dubbing solution" and "A robust alternative to **ElevenLabs**", keeps a price table for eight paid services "as of April 15, 2025", and the version 2.0 history still describes a 60-second free trial and a Shopify subscription for "unlimited usage", which no longer applies.

    AI review

Evidence is shown in its original language.

What Radar could not check about Voice-Pro

  • Whether it really runs on Mac or Linux
  • Real disk use after the roughly 10 GB of models are downloaded
  • How good the subtitles, translations and cloned voices sound in your languages
  • Whether the free Google translation endpoint keeps working without being blocked
  • Whether `browser_cookie3` is actually used to read your browser cookies and when
  • Whether the celebrity reference clips are still included in the repository you download

Similar projects

The RepoPicks app

Coming soon: install in 1 click with the RepoPicks app.

Before you download

See the release page on GitHub