OmniVoice Studio is a local, full-stack application built for fast, high-quality cinematic dubbing and custom voice generation. It wraps the 600-language zero-shot OmniVoice model into a clean workspace that works right out of the box.
- Cinematic Dubbing: Drop in an MP4. The studio transcribes the speech, automatically translates it into your chosen language, and mixes the newly generated voice back into the video.
- Background Audio Mixing: Automatically utilizes
demucsto isolate vocals. When your dubbed audio is laid over the video, the original background music and sound effects are perfectly preserved in the mix. - Voice Design & Cloning: Create entirely new voices using simple tag combinations (like
female,elderly,british accent), or instantly clone an existing voice from just a 3-second audio snippet. - Runs Locally: Fully handles its own asynchronous threading, caching, and VRAM management across Mac (Apple Silicon), NVIDIA/AMD, and CPU setups.
- Make sure
ffmpegis installed on your system. - Install Bun if you don't have it.
- Install uv if you don't have it.
- Clone and run:
git clone https://github.com/debpalash/OmniVoice-Studio.git
cd OmniVoice-Studio
uv sync
bun install
bun devThe interface will load at http://localhost:5173 (or 5174), and the API runs on port 8000. All necessary model weights will download automatically during your first generation.
Powered by the open-source OmniVoice diffusion model.