VoiceMode Voice Interface

Example Use Cases

• 1

  Add natural, conversational voice input and output to Claude Code or other MCP-compatible assistants.

• 2

  Local/offline speech processing to preserve privacy by using local STT (whisper.cpp) and TTS (Kokoro) services instead of cloud APIs.

• 3

  Real-time, low-latency multi-user voice interactions via LiveKit or local microphone/speaker transports with silence detection.

Description

VoiceMode is an MCP server that brings human-like voice interactions to Claude Code and other MCP clients. It supports local speech-to-text and text-to-speech services (e.g., whisper.cpp and Kokoro), offers silence detection to automatically stop recording, and can route audio through local devices or LiveKit rooms for real-time communication. The project is privacy-focused with options for offline/local services while allowing an OpenAI API key as a fallback. It is distributed under an MIT license and includes documentation and setup guidance for multiple platforms.