Model Context Protocol (MCP) Chat Application

This project is a fully functional Model Context Protocol (MCP) prototype built using Python, FastAPI, and Groq LLM. It supports: Dynamic context creation Versioned conversation storage Real-time streaming of assistant responses via WebSockets A web-based chat UI Environment variable management for API keys

Features

Dynamic Context Management Each chat session gets a unique context ID. Context stores the full conversation (user + assistant messages). Context can be persisted in memory (for testing) or Redis (for production). Groq LLM Integration Uses the Groq API to generate assistant responses. Supports streaming simulation token-by-token to the client. WebSocket Streaming Persistent WebSocket connection for multiple messages per session. Streams tokens in real-time to the web UI. Web Chat Interface Single-page HTML + JavaScript UI. Supports sending messages, streaming responses, and starting new chats. Dynamic context creation per session. Conversation history loads automatically when reconnecting to a context. Environment Variables API keys (e.g., Groq API key) loaded from .env using python-dotenv. Keeps secrets out of the source code.