# Monotype MCP Server & Chat Application

A complete system consisting of:
1. **MCP Server** - Plugin-ready server for Monotype API integration
2. **Backend** - Ollama-powered bridge between chat UI and MCP server
3. **Frontend** - React-based chat interface

## Architecture

```
┌─────────────┐      ┌──────────────┐      ┌─────────────┐      ┌──────────────┐
│  Frontend   │─────▶│   Backend    │─────▶│ MCP Server  │─────▶│ Monotype API │
│  (React)    │      │  (Ollama)    │      │  (Plugin)   │      │              │
└─────────────┘      └──────────────┘      └─────────────┘      └──────────────┘
```

## Project Structure

```
NextGenAgenticAI/
├── src/                    # MCP Server (can be used as plugin)
│   ├── server.js          # Main MCP server
│   ├── api-client.js      # Monotype API client
│   ├── auth.js            # Authentication service
│   ├── token-decryptor.js # Token decryption utilities
│   └── ...
├── backend/                # Backend server
│   ├── server.js          # Express server with Ollama integration
│   └── package.json
├── frontend/               # React chat UI
│   ├── src/
│   │   ├── App.jsx        # Main chat component
│   │   └── ...
│   └── package.json
└── README.md
```

## Quick Start

### 1. MCP Server (Plugin)

The MCP server can be used independently as a plugin with any chat agent.

**Setup:**
```bash
cd src
npm install
```

**Configuration:**
Add to your MCP client config:
```json
{
  "mcpServers": {
    "monotype-mcp": {
      "command": "node",
      "args": ["/path/to/src/server.js"],
      "env": {
        "MONOTYPE_TOKEN": "your-token-here"
      }
    }
  }
}
```

### 2. Backend Server

**Prerequisites:**
- Install Ollama: https://ollama.ai
- Pull llama3 model: `ollama pull llama3`

**Setup:**
```bash
cd backend
npm install
npm start
```

Server runs on `http://localhost:3001`

### 3. Frontend

**Setup:**
```bash
cd frontend
npm install
npm run dev
```

Frontend runs on `http://localhost:3000`

## Features

### MCP Server Tools

- `invite_user_for_customer` - Invite users to your company
- `get_teams_for_customer` - Get all teams
- `get_roles_for_customer` - Get all roles

### Backend Intelligence

- Uses Ollama (llama3) to detect which tool to call
- Extracts parameters from natural language
- Fallback keyword matching if Ollama unavailable

### Frontend

- Secure token input
- Modern chat interface
- Real-time responses
- Tool usage indicators

## Usage Examples

### Via Chat UI

1. Start backend and frontend
2. Enter your token
3. Try these commands:
   - "What roles are in my company?"
   - "Invite user@example.com to my company"
   - "Show me all teams"

### Via MCP Plugin

Use the MCP server directly with any MCP-compatible chat agent (like Cursor, Claude Desktop, etc.)

## Development

### Running All Services

**Terminal 1 - Backend:**
```bash
cd backend
npm run dev
```

**Terminal 2 - Frontend:**
```bash
cd frontend
npm run dev
```

**Terminal 3 - MCP Server (if testing standalone):**
```bash
cd src
npm start
```

## Environment Variables

### Backend
- `MCP_SERVER_PATH` - Path to MCP server script (default: `../src/server.js`)
- `OLLAMA_API_URL` - Ollama API URL (default: `http://localhost:11434`)

### MCP Server
- `MONOTYPE_TOKEN` - Your Monotype authentication token (optional, can be set in MCP config)

## License

MIT