⚡
Sub-second Latency
Streaming pipeline with sentence-level TTS delivers first audio in under 0.8 seconds.
EdgeVoxSub-second local voice AI
VAD > STT > LLM > TTS — fully local, fully private
Skip to content
🌍
15 Languages
English, Vietnamese, French, Korean, Thai, German, Russian, and more with language-specific STT/TTS backends.
🔒
100% Local
No cloud APIs. Whisper/Sherpa STT + Gemma LLM + Kokoro/Piper/Supertonic TTS all run on your hardware.
🎤
Voice Interrupt
Speak over the bot mid-response to cut it off for natural conversational flow.
📟
TUI + Web UI
Rich terminal UI with slash commands, plus a Web UI served via FastAPI + WebSocket.
🤖
ROS2 Ready
Optional ROS2 bridge for robotics. Publish transcriptions, subscribe to TTS commands.