RAG File MCP Server

A Model Context Protocol (MCP) server that provides Retrieval-Augmented Generation (RAG) capabilities for AI agents, with support for file-based data sources (PDF, TXT, Markdown) and a Streamlit web interface.

Features

• 🔌 MCP Server - Exposes RAG tools and resources for AI agents via the Model Context Protocol
• 📄 File Parsing - Support for PDF, TXT, Markdown, and RST files with international language detection
• 🔍 Semantic Search - Vector-based document search using configurable embedding providers
• 🎨 Streamlit UI - Modern web interface for file uploads, search testing, and log viewing
• 🐳 Docker Support - Multi-stage Docker build with docker-compose orchestration
• 📊 SQLite Logging - Built-in logging with database storage and retention policies

Quick Start

Prerequisites

• Python 3.12.7+
• Ollama (for local embeddings) or OpenAI API key

Installation

1. Clone the repository

   git clone https://github.com/pkutsch/RAGFileMcpServer.git
   cd RAGFileMcpServer

2. Create environment file

   cp .env.example .env
   # Edit .env with your configuration

3. Run the server (creates virtual environment automatically)

   # macOS/Linux
   ./run.sh
   
   # Windows
   run.bat
   
   # Cross-platform (Python)
   python run.py

4. Access the Streamlit UI

   Open http://localhost:8501 in your browser.

Configuration

Configuration is managed via environment variables. Copy .env.example to .env and customize:

Variable	Default	Description
STREAMLIT_PORT	8501	Streamlit web interface port
MCP_SERVER_PORT	8000	MCP server port (SSE mode)
EMBEDDING_PROVIDER	ollama	Embedding provider (ollama or openai)
OLLAMA_BASE_URL	http://localhost:11434	Ollama API endpoint
OLLAMA_MODEL	nomic-embed-text	Ollama embedding model
VECTOR_STORE_TYPE	chroma	Vector store (chroma, qdrant, memory)
CHUNK_SIZE	500	Text chunk size for splitting
CHUNK_OVERLAP	50	Overlap between chunks

MCP Tools

The server exposes the following tools for AI agents:

Tool	Description
search_documents	Search for documents matching a query
list_uploaded_documents	List all uploaded documents
ingest_document	Ingest a document into the RAG index
get_document_count	Get the number of indexed documents

Project Structure

RAGFileMcpServer/
├── src/
│   ├── server.py           # MCP server implementation
│   ├── streamlit_app.py    # Streamlit web interface
│   ├── file_parser/        # File parsing modules
│   │   ├── base.py         # Base parser interface
│   │   ├── pdf_parser.py   # PDF file parser
│   │   └── text_parser.py  # Text/Markdown parser
│   └── logging/            # Logging infrastructure
│       ├── db_handler.py   # SQLite log handler
│       ├── log_manager.py  # Log management
│       └── models.py       # Data models
├── tests/                  # Test suite
├── data/                   # Data directory (uploads, chroma, logs)
├── Dockerfile              # Multi-stage Docker build
├── docker-compose.yml      # Docker orchestration
├── pyproject.toml          # Python project configuration
├── run.sh / run.bat / run.py  # Cross-platform run scripts
└── .env.example            # Environment configuration template

Docker Deployment

Using Docker Compose

# Build and start services
docker-compose up -d

# View logs
docker-compose logs -f

# Stop services
docker-compose down

The compose file starts two services:

• streamlit - Web interface on port 8501
• mcp-server - MCP server (STDIO mode by default)

Standalone Docker

# Build image (run from parent directory)
docker build -f RAGFileMcpServer/Dockerfile -t rag-file-mcp-server .

# Run container
docker run -p 8501:8501 -v ./data:/app/data rag-file-mcp-server

Development

Setup Development Environment

# Create virtual environment
python -m venv .venv
source .venv/bin/activate  # or .venv\Scripts\activate on Windows

# Install with dev dependencies
pip install -e ".[dev]"

Running Tests

pytest tests/

Dependencies

This project depends on rag-core, a shared RAG module providing:

• Embedding providers (Ollama, OpenAI)
• Vector stores (ChromaDB, Qdrant, Memory)
• Text chunking strategies
• Retrieval logic

License

MIT License - see LICENSE for details.