Voicebox

Audio & Speech

jamiepine

FreeOpen SourceAPI

About

Open-source AI voice studio running locally — voice cloning, TTS across 7 engines, dictation, and agent voice in one app. A free alternative to ElevenLabs and WisprFlow, supporting 23 languages.

Visit

Share this tool

Twitter LinkedIn Facebook

Our Verdict

Highly Recommended

A new standard for open-source voice AI

Voicebox has rapidly become one of the most impressive open-source AI projects of 2026, earning 48k+ GitHub stars in just 6 months. Its combination of voice cloning, 7 TTS engines, MCP agent integration, and local-first architecture is genuinely unique — no other open-source project offers the same breadth. The single-maintainer risk and some Windows GPU issues are real concerns, but the pace of development and community response are exceptional. For anyone wanting to own their voice AI stack, Voicebox is the clear choice.

Best for

•Content creators wanting free, local voice cloning
•Developers integrating voice into AI agent workflows
•Privacy-conscious users avoiding cloud TTS services

Consider alternatives if

•You need zero-setup cloud TTS (ElevenLabs)
•You only need voice input, not output (WisprFlow)

Try Voicebox

Supported Platforms

WindowsmacOSLinuxAPI

Available platforms include Windows, macOS, Linux, and API.

Key Features

Voice cloning from as little as 3 seconds of audio

7 TTS engines: Qwen3-TTS, Qwen CustomVoice, LuxTTS, Chatterbox, TADA, Kokoro

Global hotkey dictation with Whisper STT (5 model sizes)

MCP integration for AI agents (Claude Code, Cursor, Cline)

REST API for full programmatic control

Stories Editor for multi-voice narratives

Audio effects pipeline: pitch, reverb, delay, chorus, compression

Personality system: rewrite and compose in character

23 languages supported across TTS engines

Unlimited generation length with auto-chunking and crossfade

Local or remote GPU inference (Metal, CUDA, ROCm, Intel Arc, DirectML)

End-to-end encrypted cloud backup (optional, $12/yr coming soon)

Pricing

free

Free (local app, open source, forever)

paid

Cloud $12/yr (E2E encrypted backup & sync, coming soon)

Use Cases

Content creators needing voiceovers and narration

AI agent voice output for development workflows

Podcast and audiobook production with multi-voice support

Game NPC dialogue generation and localization

Pros

Completely free and open source (MIT license)

Truly local-first — everything runs on your machine, no cloud dependency

Combines voice cloning, TTS, dictation, and agent voice in one app

Active development with 48k+ GitHub stars and responsive maintainer

MCP integration makes it a natural fit for AI agent workflows

7 TTS engines with 23 languages offer exceptional flexibility

Cons

No pre-built Linux binaries (must build from source)

Windows GPU support has issues in prebuilt binaries

572 open issues on GitHub — many feature requests but some bugs

Cloud backup and sync not yet available (coming soon)

Single maintainer-driven — bus factor risk

Requires GPU for optimal performance (CPU works but slower)

Latest Update

2026-08: v0.5.0 released, 48k+ GitHub stars, 1.5M+ downloads

Get Started in 4 Steps

Get started in 5 steps

Download Voicebox

Download the latest release from GitHub for your platform (macOS, Windows, Linux).

Download

https://github.com/jamiepine/voicebox/releases

Clone or select a voice

Import an audio file, record from mic, or capture system audio. Voicebox clones from as little as 3 seconds.

Generate speech

Type or paste text and choose a TTS engine. Voicebox supports 7 engines with different strengths.

API

curl -X POST http://127.0.0.1:17493/generate -H "Content-Type: application/json" -d '{"text":"Hello world","profile_id":"YOUR_PROFILE_ID","engine":"qwen_custom_voice"}' --output hello.wav

Connect to AI agents via MCP

Add Voicebox to your MCP config so AI agents can speak in your cloned voices.

MCP Config

{
  "mcpServers": {
    "voicebox": {
      "url": "http://127.0.0.1:17493/mcp"
    }
  }
}

Related Audio & Speech Tools

ElevenLabs

Leading AI voice synthesis and cloning platform with multi-language support

Whisper

OpenAI

OpenAI's open-source speech recognition model for multi-language speech-to-text

Fish Audio

Voice generation platform from the team behind the open-source TTS star Fish Speech. Clone a voice from just 10-30 seconds of audio; the S1/S2 models deliver natural, expressive speech with commercial use and pay-as-you-go API

MiniMax Audio

MiniMax

MiniMax's voice generation platform. The Speech model family delivers hyper-realistic TTS in 40+ languages with 10-second voice cloning, controllable emotion and sound-effect tags, and a free web trial