Maid

A full-featured MCP server that controls a maid Mimi with voice, avatar, and speech recognition for Claude Desktop.
  • javascript

1

GitHub Stars

javascript

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "quinny1187-maid-mcp": {
      "command": "node",
      "args": [
        "path/to/maid-mcp/maid-server.js"
      ]
    }
  }
}

You can run a full MCP server that gives Claude Desktop a maid personality named Mimi, complete with voice, avatar, and speech recognition. It integrates tightly with Claude Desktop to provide interactive tools for speaking, avatar control, and responsive interactions, all managed through a single Node.js server.

How to use

Install and configure the MCP server, then connect it to your MCP client to start interacting with Mimi. Use the provided tools to make Mimi speak, change voice, animate the avatar, and respond to your voice commands. The server runs locally and communicates with Claude Desktop to process input, generate responses, and render avatar animations and audio.

How to install

# 1) Install Node.js dependencies (if not already done)
npm install

# 2) Install Python dependencies for voice input
cd voice
install_voice_deps.bat
cd ..

# 3) Install Python dependencies for avatar (if needed)
c d avatar
install_avatar_deps.bat
cd ..
# 2) Configure Claude Desktop by adding the MCP server config
# This block is a snippet you paste into your setup to register the MCP server
{
  "mcpServers": {
    "maid": {
      "command": "node",
      "args": ["path/to/maid-mcp/maid-server.js"]
    }
  }
}
# 3) Launch Everything

# Recommended: Use Python launcher for best process management
start_all_python.bat

# Alternative: Use enhanced batch launcher
start_all.bat
# 4) Stop Everything
stop_all.bat

Additional setup and flow

This setup automatically cleans up existing processes, shows the avatar, starts the avatar state server on port 3338, and opens the voice input listener. You can then speak to Mimi, have her respond, and watch the avatar react with animations.

Voice Loop

You speak into the microphone, speech recognition converts your words to text, the ultra-fast sender forwards the text to Claude Desktop, Mimi processes the input, then Mimi speaks with a Japanese-accented voice and the avatar animates in response.

Avatar interaction and controls

Use the interactive controls to show, hide, or move the avatar, and to trigger animations. Right-click hides the avatar, double-click closes it, left-click cancels an animation, and drag moves the avatar while showing a pick-up pose. Pressing ESC closes the avatar permanently.

Available tools

speak

Convert text to speech with a selectable emotion (neutral, happy, sad, excited, angry, shy) and handle voice output via the maid system.

list_voices

List available TTS voices that can be used by Mimi.

set_voice

Change the current TTS voice by voiceId.

show_avatar

Display the interactive Mimi avatar on screen with optional animation and position parameters.

hide_avatar

Hide the avatar while keeping the system running.

play_animation

Play a specific animation or pose by its id.

stop_animation

Stop the currently running animation.

move_avatar

Reposition the avatar using x and y coordinates.

create_animation

Create a custom sequence with an id, name, frames, fps, and loop flag.

list_animations

List all available animations.

list_poses

List all available avatar poses.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational
Maid MCP Server - quinny1187/maid-mcp | VeilStrat