Modern Control Protocol (MCP) Server

A modern, scalable MCP server implementation with support for multiple AI providers, advanced monitoring, and robust conversation management.

Features

• Multi-provider AI support (OpenAI, Anthropic, Google AI, Azure)
• Real-time streaming responses
• Conversation management and history
• Function calling and tool usage
• Vector database integration
• Semantic caching
• Prometheus metrics and Grafana dashboards
• Rate limiting and error handling
• PostgreSQL for data persistence
• Redis for caching
• Elasticsearch for search
• Docker containerization

Quick Start

1. Clone the repository
2. Copy environment template:

   cp .env.example .env
   

3. Update environment variables in .env
4. Start services with Docker Compose:

   docker-compose up -d
   

API Documentation

Once running, visit:

• API Documentation: http://localhost:8000/docs
• ReDoc Documentation: http://localhost:8000/redoc

Monitoring

• Prometheus metrics: http://localhost:9090
• Grafana dashboards: http://localhost:3000

Development

Prerequisites

• Python 3.9+
• PostgreSQL
• Redis
• Elasticsearch
• Docker & Docker Compose

Local Setup

1. Create virtual environment:

   python -m venv venv
   source venv/bin/activate  # Linux/Mac
   # or
   .\venv\Scripts\activate  # Windows
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Run development server:

   uvicorn app.main:app --reload
   

Testing

Run tests with:

pytest

Architecture

The MCP server is built with a microservices architecture:

• FastAPI for the REST API
• PostgreSQL for data persistence
• Redis for caching and rate limiting
• Elasticsearch for search functionality
• Qdrant for vector storage
• Prometheus and Grafana for monitoring

API Endpoints

• /api/v1/mcp/prompts: Prompt management
• /api/v1/mcp/conversations: Conversation handling
• /api/v1/mcp/conversations/{id}/complete: AI completions
• /metrics: Prometheus metrics
• /health: Health check

License

MIT License