MCP RAG Chatbot

A powerful Model Context Protocol (MCP) based Retrieval-Augmented Generation (RAG) chatbot that intelligently searches through your document knowledge base to provide accurate, context-aware responses. Built with GroundX for document storage and retrieval, and Google's Gemini AI for natural language generation.

Features

• 🔍 Intelligent Document Search: Leverages GroundX's powerful semantic search capabilities
• 🤖 AI-Powered Responses: Uses Google Gemini 2.0 Flash for generating contextual answers
• 📚 Document Ingestion: Easy document upload and processing (PDF, TXT, DOCX, etc.)
• 🌐 Web Interface: Beautiful, responsive chat interface with real-time messaging
• 🔌 MCP Integration: Standard Model Context Protocol implementation for tool interoperability
• 🚀 RESTful API: Complete HTTP API for programmatic access
• 📊 Health Monitoring: Built-in health checks and comprehensive logging
• ⚡ Real-time Processing: Fast document search with context-aware responses

Architecture

┌─────────────────┐    ┌──────────────┐    ┌─────────────────┐
│   Web Client    │◄───┤ Flask Server ├───►│   MCP Server    │
└─────────────────┘    └──────────────┘    └─────────────────┘
                              │                       │
                              ▼                       ▼
                       ┌──────────────┐    ┌─────────────────┐
                       │   Gemini AI  │    │    GroundX      │
                       │ (Text Gen.)  │    │ (Doc Storage)   │
                       └──────────────┘    └─────────────────┘

Quick Start

Prerequisites

• Python 3.8+
• GroundX API account
• Google AI Studio API key

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/mcp-rag-chatbot.git
   cd mcp-rag-chatbot

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables:

   cp .env.example .env
   # Edit .env with your API keys

4. Configure your .env file:

   GROUNDX_API_KEY=your_groundx_api_key_here
   GEMINI_API_KEY=your_gemini_api_key_here

Running the Server

python newserver.py

The server will start with:

• Web Interface: http://localhost:8080
• MCP Server: Running on stdio
• API Endpoints: Available at base URL

Usage

Web Interface

1. Open http://localhost:8080 in your browser
2. Start asking questions about your documents
3. The system will automatically search for relevant context and provide informed answers

MCP Integration

The server implements standard MCP tools that can be used by any MCP-compatible client:

• search_doc_for_rag_context(query: str) - Search documents for relevant context
• ingest_documents(local_file_path: str, file_type: str) - Add documents to knowledge base

Configuration

Environment Variables

Variable	Description	Default
GROUNDX_API_KEY	GroundX API key (required)
GEMINI_API_KEY	Google AI Studio API key (required)

API Reference

Endpoints

GET /

Serves the main chat web interface.

POST /chat

Main RAG endpoint for conversational queries.

Request:

{
  "query": "Your question here"
}

Response:

{
  "answer": "AI-generated response",
  "context_used": true,
  "query": "Your original question"
}

Dependencies

• fastmcp: MCP server implementation
• groundx: Document storage and retrieval
• google-generativeai: Gemini AI integration
• flask: Web server framework
• flask-cors: CORS support
• python-dotenv: Environment variable management

How It Works

1. Document Ingestion: Upload documents to GroundX buckets for semantic indexing
2. Query Processing: User questions are processed and relevant context is retrieved
3. Context Augmentation: Retrieved context is combined with the user query
4. AI Generation: Gemini AI generates informed responses based on the context
5. Response Delivery: Contextual answers are delivered via web interface or API

Troubleshooting

Common Issues

"GROUNDX_API_KEY environment variable is not set"

• Ensure your .env file contains the correct API key
• Verify the .env file is in the same directory as newserver.py

"No relevant documents found"

• Make sure documents have been ingested into your GroundX bucket
• Wait a few minutes after ingestion for documents to be indexed
• Try different query phrasings

Connection errors

• Check your internet connection
• Verify API keys are valid and have sufficient quotas

Logs

The server provides comprehensive logging. Check the console output for detailed information about:

• Document search operations
• AI response generation
• API connectivity status
• Error details