# jamiepine/voicebox

The open-source AI voice studio. Clone, dictate, create.

Repository: https://github.com/jamiepine/voicebox
Canonical: https://ross.abutalabs.com/products/voicebox
Homepage: https://voicebox.sh
Language: TypeScript
License: MIT
License Family: permissive
Topics: ai, voice-clone, qwen3-tts, voice-ai, whisper, qwen3-tts-ui, cuda, mlx
Last push: 2026-08-09T00:03:42+00:00

## Health v2 (maintenance only)
Score: 75/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 96, release rhythm 81, longevity 15
- inputs: {"age_days": 220, "days_push": 25, "days_rel": 130, "gap_med": 0.0, "n_releases_24m": 25}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 51551, forks 6430 (observed 2026-08-28T04:12:11.876917+00:00)

## What it is
Voicebox is a free, open-source, local-first AI voice studio desktop app that combines voice cloning, text-to-speech across 7 engines, and system-wide voice dictation. It runs entirely on your own hardware and can give MCP-aware AI agents a voice of your choosing.

## Use cases
- clone a voice from a few seconds of audio
- generate speech locally instead of paying for ElevenLabs
- dictate into any app with a global hotkey instead of WisprFlow
- give my AI agent a custom voice via MCP
- generate multilingual TTS in 23 languages
- create multi-voice audio projects and stories
- transcribe and archive voice recordings locally

## When to choose
- you want free, private, fully local voice cloning and TTS with no account or subscription
- you need both speech output (TTS) and input (dictation) in one desktop app
- you want to give MCP-aware AI agents a voice you own
- you have a GPU (CUDA or Apple Silicon) and want offline speech generation

## When to avoid
- you need a managed cloud API with guaranteed uptime and no local hardware
- you need a prebuilt Linux binary — Linux currently requires building from source
- you only need lightweight command-line TTS scripting rather than a GUI studio
- you require commercial licensing guarantees for cloned celebrity-style voices

## Facets
- artifact type: application
- maturity: active
- function: tts, speech-recognition, llm-inference, mcp, audio-processing, gui
- domain: speech-processing, artificial-intelligence, desktop-applications, self-hosted
- platform: windows, cross-platform
- tags: voice-cloning, dictation, local-first, tauri, elevenlabs-alternative, wisprflow-alternative, qwen3-tts, whisper, voice-studio, mcp-server, audio, macos, linux, desktop, gpu

## Member repositories
- jamiepine/voicebox (main) score 75

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:11.876917+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:19:45.456008+00:00, confidence not recorded.
  - readme: https://github.com/jamiepine/voicebox (fetched 2026-08-28T04:12:11.876917+00:00, sha 7d37a432c73c)
  - homepage: https://voicebox.sh (fetched 2026-08-28T18:09:31.328173+00:00, sha 7b390bf104c5)
  - site_page: https://voicebox.sh/linux-install (fetched 2026-08-28T18:09:31.340902+00:00, sha 62d404dabb07)
  - site_page: https://docs.voicebox.sh (fetched 2026-08-28T18:09:31.339195+00:00, sha f6385afe3cdb)
  - site_page: https://voicebox.sh/pricing (fetched 2026-08-28T18:09:31.337333+00:00, sha 786a7fbe36b2)
  - site_page: https://voicebox.sh/token (fetched 2026-08-28T18:09:31.342406+00:00, sha a2cbf64af9ca)
- Data as of 2026-08-30T08:39:29.467469+00:00.