voice-pro
An AI-powered web app for speech recognition, translation, and zero-shot voice cloning.
Multilingual dubbing, transcription, and zero-shot voice cloning in a single Gradio interface.
Voice-Pro integrates YouTube downloading, Demucs vocal separation, Whisper speech recognition, and multiple TTS engines into a local web application. It is designed for creators, researchers, and developers who need advanced audio dubbing and translation tools without relying entirely on paid SaaS APIs.
What it is
The application provides a Dubbing Studio tab to pull audio from YouTube, strip background noise with Demucs, and generate subtitles across more than 100 languages. It supports multiple transcription backends like Faster-Whisper alongside zero-shot voice cloning frameworks including F5-TTS, E2-TTS, and CosyVoice.
Who it's for
It suits content creators, podcasters, and researchers working on Windows who want localized control over translation, transcription, and voice generation pipelines. It does not suit users looking for an out-of-the-box cloud service, Mac-native binaries, or minimal setup requirements.
Key features
Zero-Shot Voice Cloning
Clone voices using F5-TTS, E2-TTS, and CosyVoice models with multi-language support.
Dubbing Studio Workflow
Download YouTube videos, isolate vocals with Demucs, transcribe, translate, and generate TTS in one hub.
Multiple ASR Engines
Process speech using Whisper, Faster-Whisper, and Whisper-Timestamped models for accurate subtitles.
Corporate PC Friendliness
Auto-downloads portable ffmpeg and features self-healing model downloads and network rate-limit retries.
Up and running in 3 steps
Our honest verdict
What's great
- Combines YouTube extraction, Demucs vocal isolation, transcription, and translation into a unified Gradio interface.
- Supports multiple top-tier speech synthesis and cloning engines including F5-TTS, Kokoro, and CosyVoice.
- Fast installer migration using uv with prebuilt wheels that remove the need for manual CUDA Toolkit installations.
- Completely free and open source with active community support via GitHub issues and discussions.
Where it falls short
- Operation on macOS and Linux has not been verified, requiring a Windows environment with an NVIDIA GPU for reliable performance.
- Initial runs require downloading large AI models such as CosyVoice2-0.5B, which can take over an hour depending on network speed.
- Development is currently paused by the original maintainers, shifting long-term maintenance entirely to the community.
Pricing & license
The software is completely free and open source under its GPL-3.0 licence. The only cost is running it yourself on local hardware with sufficient GPU resources.
Open remote AI-training roles
874 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
Use it, or have us set it up
FAQ
What operating systems are supported?
It works well on Windows with an NVIDIA GPU. Operation on Mac and Linux has not been verified.
Are CUDA Toolkit and Visual Studio Build Tools required?
No, all dependencies ship as prebuilt wheels and PyTorch bundles the CUDA runtime.
Can I use commercial APIs with Voice-Pro?
Yes, you can optionally configure your own Microsoft Azure keys for TTS and translation services.
