Local Voice Mcp
Give Claude Code or Cursor voice output without sending audio to a cloud TTS API by running Chatterbox locally through MCP.
Overview
Local Voice MCP is a Build-phase MCP server that lets MCP clients speak responses using local Chatterbox TTS instead of a cloud voice API.
What is this MCP server?
- stdio MCP server `@codecraftersllc/local-voice-mcp` v0.1.5 from npm
- Runs Chatterbox TTS on your machine so assistants can speak responses locally
- No cloud voice vendor required for basic text-to-speech from MCP clients
- Fits solo builders who want private, offline-capable agent feedback
- Published MCP server version 0.1.5
- npm package identifier @codecraftersllc/local-voice-mcp
- Transport type stdio per server schema
What problem does it solve?
Cloud TTS adds cost, latency, and data leaving your machine when you only want your coding agent to read answers aloud.
Who is it for?
Solo builders extending Claude Code or Cursor with offline-friendly local TTS for demos or accessibility.
Skip if: Teams that need managed voice SLAs, telephony, or multi-user voice products without running local ML.
What do I get? / Deliverables
After you register the stdio server, your MCP client can trigger on-device Chatterbox speech for assistant output.
- Registered stdio MCP server entry for local voice
- Speak-capable tool surface for connected MCP clients
- On-device synthesized audio from agent-triggered text
Recommended MCP Servers
Journey fit
Voice capability is wired when you extend the agent runtime during the build phase, not when you ship or grow the product. MCP stdio servers that add modalities to the coding agent belong on the agent-tooling shelf alongside other MCP integrations.
How it compares
MCP TTS integration, not a general brainstorming or planning skill.
Common Questions / FAQ
Who is local-voice-mcp for?
Indie developers and agent users who want MCP-connected assistants to speak using local Chatterbox models on their own machine.
When should I use local-voice-mcp?
Use it during build when you are configuring MCP servers and want voice feedback without a cloud TTS subscription.
How do I add local-voice-mcp to my agent?
Install `@codecraftersllc/local-voice-mcp` from npm, add a stdio MCP entry in Claude Code or Cursor pointing at that package, and restart the client.