Tool not found.
Back to Home
Local-first, open-source AI voice studio, clone any voice, generate speech in 23 languages, dictate into any app, and give AI agents a voice, all running on your machine.
Voicebox is a free, open-source alternative to ElevenLabs and WisprFlow combined into a single local app. It covers the full voice I/O loop: speech output via seven TTS engines and speech input via Whisper-based dictation with a global hotkey. Voice cloning works from a few seconds of reference audio with zero-shot cloning, or you can pick from 50+ curated preset voices. Supported engines include Qwen3-TTS, Qwen CustomVoice, LuxTTS, Chatterbox Multilingual, Chatterbox Turbo, HumeAI TADA, and Kokoro, covering 23 languages. Post-processing effects include pitch shift, reverb, delay, chorus, compression, and filters. A Stories editor provides a multi-track timeline for conversations, podcasts, and narratives. MCP-aware AI agents like Claude Code and Cursor can speak to you in a cloned voice via a single tool call. A bundled local Qwen3 LLM handles dictation refinement and voice personality rewrites entirely on-device. Voicebox is built with Tauri and FastAPI, runs on macOS, Windows, Linux, and Docker, and supports NVIDIA CUDA, AMD ROCm, Apple Silicon MLX, Intel Arc, and CPU. All models, voice data, and captures stay on your machine — nothing leaves your device.
