What it does

VoiceStudio is an open-source, fully local voice tool. Its workspaces cover cloning a voice or designing one, dubbing video with timed speech, dictating through a floating widget, producing stories and audiobooks as batch jobs, and reusable recipes with watch folders. The site cites support for 646 languages. The default engine is VoiceStudio itself, powered by k2-fsa/OmniVoice, and you can choose other engines from its catalog. To clone a voice you pick a voice or add a clean reference recording, enter your text and generate, installing the required model when prompted. Local workflows run on your hardware; remote services are optional and usage analytics requires consent. A local API and an MCP server let agents use it too. It ships as an Electron desktop app with installers for macOS, Windows and Linux, plus a one-command install script. It uses CUDA on NVIDIA GPUs and Metal on Apple Silicon, and can run on a CPU more slowly. iOS and Android phone companion apps are listed as a preview. The application is licensed AGPL-3.0. Model weights, tokenizers, voices and generated-output conditions are separate, and the site asks users to clone voices only with permission.

Capabilities

Voice cloning from a reference recording, and voice design
Video dubbing with timed speech
Dictation through a floating widget, plus transcription
Stories, audiobooks and batch jobs with reusable recipes and watch folders
Local API and MCP server for agents
Runs locally with CUDA, Apple Silicon or CPU support

When to use it

Cloning or designing voices without sending audio to a cloud service
Dubbing a video into another language on your own hardware
Producing audiobooks or batches of narration locally
Dictating text or transcribing recordings on your own machine

Alternatives to explore

ElevenLabs