VoiceStudio (OmniVoice): Windows cloning and dubbing

This tutorial covers installing OmniVoice Studio on Windows, then using it for voice cloning, speech generation and video dubbing. The presenter describes it as an open source alternative to ElevenLabs that processes audio locally on a PC. Audiobook creation appears in the introduction, while the practical walkthrough focuses on setup and short audio and video workflows.

The stated requirements include a graphics card with at least 6 GB, Python 3.10, FFmpeg, Visual Studio and Bun. The presenter also names toolkit versions 11.8 to 12.8 and checks CUDA before installing PyTorch. Setup proceeds through cloning the repository, creating and activating a Python virtual environment, installing PyTorch with torchaudio, and updating Python packaging tools. Bun installation and UV sync follow. After launch, the walkthrough downloads models before entering the studio. These are the video's setup instructions, rather than an independently verified compatibility guide.

For voice cloning, the presenter uploads an audio file, names and saves a voice profile, then enters text and synthesizes speech with a selected voice. A separate voice-design example uses text and a speaker selection. The dubbing workflow starts with a video upload and transcription, followed by language selection, translation, dub generation and export. The supplied description says to use only voices for which you have permission.