Voicebox: A Local-First Open-Source ElevenLabs Alternative With a REST API for Agents
GitHub·medium signal
Voicebox, from Spacedrive creator Jamie Pine, is a local-first, MIT-licensed AI voice studio (~28.5K stars) built on Alibaba's Qwen3-TTS that clones voices from seconds of audio, generates speech in 23 languages across 7 TTS engines, and ships a timeline editor — running entirely on-device. Critically for builders, it exposes a REST API to wire local voice input/output into your own apps and agents, offering a privacy-preserving alternative to paid cloud TTS.