voicebox

An open-source local voice studio for cloning voices, generating speech and dictating into any app. It runs on your own machine.

Share on XLicense: MIT

Overview

Voicebox is a local-first, open-source voice studio that runs on your own machine. It clones voices from a few seconds of audio, generates speech, and dictates into any text field with a global hotkey. It supports 23 languages across 7 TTS engines and can give MCP-aware AI agents a chosen voice. A bundled local LLM handles refinement and per-profile personas.

Key features

  • Voice cloning from a few seconds of audio
  • Speech generation in 23 languages across 7 TTS engines
  • Dictation into any text field with a global hotkey
  • Gives MCP-aware AI agents a voice you choose
  • Models and voice data stay on your machine

Best for

Users who want voice generation and dictation without a cloud service, in one desktop app.

Upstream
jamiepine/voicebox
Fork on GitHub
Guo-astro/voicebox
Upstream stars
56k
Category
Documents, media and content
Language
TypeScript
License
MIT
Forked
2026-02-18
Sync status
In syncLast synced 2026-09-29