Mai Buddy

Mai Buddy is a highly customizable, personal AI agent

Comment

Mai Buddy Desktop AI Agent

Mai Buddy is a highly customizable, personal AI agent that runs as a desktop application. It features GPT integration, ElevenLabs voice synthesis, and dozens of MCP (Model Context Protocol) connections for extending functionality.

Features

AI Integration

  • GPT-4 Support: Powered by OpenAI's latest models
  • Customizable System Prompts: Tailor the AI's personality and behavior
  • Conversation Memory: Maintains context across sessions
  • Multiple Model Support: Choose from GPT-4, GPT-4 Turbo, and GPT-3.5 Turbo

Voice Capabilities

  • ElevenLabs Integration: High-quality text-to-speech
  • Voice Activation: Respond to "Mai Buddy" trigger phrase
  • Multiple Voice Options: Choose from various AI voices
  • Real-time Speech Recognition: Voice input support

MCP Connections (30+ Integrations)

  • Development: GitHub, GitLab, VS Code, Docker
  • Communication: Slack, Discord, Microsoft Teams
  • Productivity: Notion, Trello, Asana
  • Cloud Services: AWS, Google Cloud, Azure, DigitalOcean
  • Databases: PostgreSQL, MySQL, MongoDB, Redis
  • AI/ML: Hugging Face, Anthropic, Replicate
  • Finance: Stripe, Plaid, Salesforce
  • Media: YouTube, Spotify, Unsplash
  • System: File System, Terminal, Calendar
  • Custom: Build your own MCP servers

User Experience

  • Always Available: Runs in system tray
  • Global Hotkeys: Quick access with keyboard shortcuts
  • Customizable Interface: Modern, responsive design
  • Cross-Platform: Works on macOS, Windows, and Linux

Installation

Prerequisites

  • Node.js 18+
  • npm or yarn
  • OpenAI API key
  • ElevenLabs API key (optional, for voice features)

Quick Start

  1. Clone the repository

    git clone https://github.com/djleamen/mai-buddy.git
cd mai-buddy

  2. Install dependencies

    npm install

  3. Start the application

    npm start

  4. Configure API keys

    • Click the settings icon in the app
    • Add your OpenAI API key
    • Optionally add ElevenLabs API key for voice features

Configuration

API Keys

  • OpenAI: Required for AI functionality
  • ElevenLabs: Optional for text-to-speech
  • Service-specific keys: Required for each MCP connection you want to use

Hotkeys (Default)

  • Cmd/Ctrl + Shift + M: Show/hide chat window
  • Cmd/Ctrl + Shift + V: Voice activation
  • Cmd/Ctrl + Shift + C: Quick capture
  • Cmd/Ctrl + Shift + L: Toggle listening
  • Escape: Hide window

Voice Settings

  • Voice Selection: Choose from available ElevenLabs voices
  • Stability: Control voice consistency (0.0 - 1.0)
  • Similarity Boost: Enhance voice similarity (0.0 - 1.0)
  • Trigger Phrase: "Mai Buddy" (customizable)

MCP Connections

Mai Buddy supports the Model Context Protocol for extending functionality. Available connection types:

Development Tools

  • GitHub: Repository management, issue tracking, code search
  • GitLab: Project management, CI/CD, merge requests
  • VS Code: Editor integration, file operations
  • Docker: Container management, image building

Communication Platforms

  • Slack: Team messaging, channel management
  • Discord: Server management, voice channels
  • Microsoft Teams: Collaboration, meetings

Productivity Services

  • Notion: Database management, page creation
  • Trello: Board management, card operations
  • Asana: Project tracking, task management

Cloud Platforms

  • AWS: EC2, S3, Lambda management
  • Google Cloud: Compute Engine, Cloud Storage
  • Azure: Virtual machines, storage accounts
  • DigitalOcean: Droplet management, Spaces

Databases

  • PostgreSQL: Query execution, schema management
  • MySQL: Table operations, data import
  • MongoDB: Document operations, collections
  • Redis: Key-value operations, caching

Development

Building

# Development mode
npm run dev

# Build for current platform
npm run build

# Build for specific platforms
npm run build-mac
npm run build-win
npm run build-linux

Custom MCP Servers

Create custom MCP servers to extend Mai Buddy's capabilities:

// Example custom MCP server
const WebSocket = require('ws');

const server = new WebSocket.Server({ port: 3001 });

server.on('connection', (ws) => {
  ws.on('message', (data) => {
    const message = JSON.parse(data);

    if (message.method === 'tools/call') {
      // Handle custom tool calls
      const result = handleCustomTool(message.params);

      ws.send(JSON.stringify({
        jsonrpc: '2.0',
        id: message.id,
        result: result
      }));
    }
  });
});

Mai Buddy - Your personal AI companion, always ready to assist! 🤖✨

Built With

Share this project:

Updates