MCP Platform

High-performance Model Context Protocol (MCP) client with WebSocket API, SQLite storage, and support for OpenRouter, OpenAI, and Groq.

🚀 Quick Start

# Install and run
uv sync
export OPENROUTER_API_KEY="your_key_here"  # or OPENAI_API_KEY, GROQ_API_KEY
./run.sh

Connect: ws://localhost:8000/ws/chat

📡 WebSocket API

Send:

{
  "action": "chat",
  "request_id": "unique-id",
  "payload": {
    "text": "Hello",
    "streaming": true
  }
}

Receive:

{
  "request_id": "unique-id",
  "status": "chunk",
  "chunk": {
    "type": "text",
    "data": "Response...",
    "metadata": {}
  }
}

⚙️ Configuration

LLM Provider (src/config.yaml)

llm:
  active: "openrouter"  # or "openai", "groq"
  providers:
    openrouter:
      model: "openai/gpt-4o-mini"
    openai:
      model: "gpt-4o-mini"
    groq:
      model: "llama-3.3-70b-versatile"

MCP Servers (src/servers_config.json)

{
  "mcpServers": {
    "my_server": {
      "enabled": true,
      "command": "python",
      "args": ["path/to/server.py"]
    }
  }
}

📁 Key Files

• src/config.yaml - LLM providers and settings
• src/servers_config.json - MCP server configurations
• chat_history.db - SQLite database for conversation history

✅ Features

• Full MCP Protocol - Tools, prompts, resources
• High Performance - SQLite with WAL mode, optimized indexes
• Real-time Streaming - WebSocket with delta persistence
• Multi-Provider - OpenRouter (100+ models), OpenAI, Groq
• Type Safe - Pydantic validation throughout

---

Requirements: Python 3.13+, request_id required in WebSocket messages