Log inSubmit server
Voice MCP logo

Voice MCP

by mbailey

Adds microphone-based voice conversations to Claude Code and MCP agents with local or cloud speech services.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Voice MCP

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "mbailey-voice-mcp": {
      "command": "uvx",
      "args": [
        "voice-mode-install"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

mbailey/voice-mcp connects MCP-capable agents to a computer’s microphone and speakers for spoken conversations. It supports speech-to-text and text-to-speech through optional local Whisper.cpp and Kokoro services or OpenAI-compatible cloud services, with automatic silence detection for recording. Reach for it when an agent needs hands-free interaction, offline voice processing, or a break from keyboard input.

Use cases

  • Start hands-free conversations with Claude Code
  • Use voice input while cooking, walking, or holding objects
  • Run speech processing locally for offline use
  • Switch to OpenAI-compatible cloud speech services
  • Debug microphone and audio output problems with saved recordings

Key features

  • Speech-to-text through Whisper.cpp or OpenAI-compatible services
  • Text-to-speech through Kokoro or OpenAI-compatible services
  • Local microphone capture and speaker playback
  • Automatic silence detection
  • Claude Code plugin and MCP launcher setup
  • Linux, macOS, Windows, WSL, and NixOS support

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

18/18 checks healthy over the last 78d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 9d ago via OSV.dev · voice-mode-install (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars1.4kStargazers on the source repository.
Last commit1d agoMost recent push to the default branch.
Directory activity5 viewsConfig copies, upvotes, and views on AllMCPs.

How Voice MCP works

What mbailey/voice-mcp does

mbailey/voice-mcp adds spoken interaction to Claude Code and other agents that support MCP. It captures speech from a local microphone, sends it through a speech-to-text service, passes the resulting conversation through the connected agent, and plays the response through the computer’s speakers.

The project is designed for situations where typing is inconvenient or visual attention is limited. It can use cloud services through an OpenAI API key, or run speech processing locally with Whisper.cpp for transcription and Kokoro for speech synthesis. Local services provide an OpenAI-compatible interface, allowing the voice server to switch between local and cloud processing.

How it works

VoiceMode runs as a Python package and exposes its functionality through an MCP launcher. Claude Code can install it as a plugin or register it with claude mcp add. The plugin provides commands for installing the required dependencies and starting a conversation.

During a conversation, recording ends when the silence detector determines that the speaker has stopped. The agent response is then converted to audio and played locally. A microphone and speakers are required for the basic setup.

The server supports Linux, macOS, Windows, Windows Subsystem for Linux, and NixOS. The documented Python compatibility range is 3.10 through 3.14. Platform-specific packages may be needed for audio capture and playback, including PortAudio, FFmpeg, PulseAudio components, or development libraries.

Setup and configuration

Claude Code users can install the marketplace plugin, install the VoiceMode plugin, run its dependency installer, and start the conversation command. An alternative installs the Python package and dependencies with uv, then registers voicemode-mcp-launcher as a user-scoped MCP server.

Set OPENAI_API_KEY when using OpenAI cloud services or as a fallback for local services. Configuration can also be edited with voicemode config edit. Local-only setups can install Whisper.cpp and Kokoro instead of relying on cloud speech APIs.

For troubleshooting, check application permissions when the microphone is unavailable and verify audio device settings when playback fails. WSL2 users need PulseAudio packages for microphone access. Setting VOICEMODE_SAVE_AUDIO=true saves recordings under ~/.voicemode/audio/YYYY/MM/ for debugging.

Tools and capabilities

  • Start natural voice conversations through the converse MCP tool.
  • Manage or interact with voice services through the service MCP tool.
  • Transcribe speech with local Whisper.cpp or an OpenAI-compatible service.
  • Generate spoken responses with local Kokoro or an OpenAI-compatible service.
  • Detect conversational silence and stop recording automatically.
  • Choose between local processing and cloud services for privacy, connectivity, or operational needs.

The example Claude Code permissions use mcp__voicemode__converse and mcp__voicemode__service, matching the two exposed MCP operations.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Voice MCP

mbailey/voice-mcp is an MCP server that connects Claude Code and other MCP-capable agents to a local microphone and speakers. Its main tools are `converse` for spoken conversations and `service` for voice-service operations, using local Whisper.cpp and Kokoro or OpenAI-compatible speech APIs.

Install it through the Claude Code plugin marketplace or with the Python installer package. The plugin path installs VoiceMode and its dependencies; the Python path uses `uvx voice-mode-install`, then registers `voicemode-mcp-launcher` with `claude mcp add`.

Yes, it can work offline when local Whisper.cpp speech-to-text and Kokoro text-to-speech services are installed. These local services replace cloud speech processing, while OpenAI-compatible APIs remain available when cloud access is preferred.

Set `OPENAI_API_KEY` for OpenAI cloud services or use `voicemode config edit` for configuration changes. Local speech services can be installed separately, and `VOICEMODE_SAVE_AUDIO=true` enables audio capture for debugging.

Related MCP servers

All in Conversational AI All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewVoice MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/mbailey-voice-mcp?style=directory)](https://allmcps.com/mcp/mbailey-voice-mcp)
HTML Embed
<a href="https://allmcps.com/mcp/mbailey-voice-mcp"><img src="https://allmcps.com/api/badge/mbailey-voice-mcp?style=directory" alt="Voice MCP on AllMCPs" /></a>