Voicebox
A local-first AI voice studio that clones voices, synthesizes speech through seven TTS engines, provides system-wide dictation, and connects MCP agents to custom voices.
Voicebox combines voice input and output in one desktop application: Whisper-based speech-to-text with a global dictation hotkey that pastes into the focused field, and text-to-speech through seven engines, including Qwen3-TTS, Qwen CustomVoice, LuxTTS, Chatterbox and HumeAI TADA, with zero-shot voice cloning from a short reference sample and effects such as pitch, reverb and compression. All inference runs on the user's own hardware, and audio does not leave the machine unless remote access is explicitly configured.
An HTTP MCP server exposes its voice capabilities to MCP agents for agent-driven voice workflows, and a bundled local Qwen3 model refines dictated transcripts and drives per-profile personas. Synthesis supports 23 languages.
The desktop app is built with Tauri and Rust, sharing a React UI with a browser-accessible web build, and is written in TypeScript.