A powerful documentation assistant that combines Model Context Protocol (MCP) with Claude AI to provide intelligent Q&A over your documentation files. Built with FastMCP and Gradio for an easy-to-use web interface.

• 🔍 Smart Document Search: Full-text search across your documentation files
• 🤖 AI-Powered Responses: Uses Claude AI to provide intelligent answers based on your docs
• 📁 Multi-Format Support: Works with .md, .txt, .rst, .pdf, and image files
• 🔤 OCR Capabilities: Extract text from image-based PDFs and standalone images using Tesseract OCR
• 📄 Advanced PDF Support: Hybrid PDF processing with automatic fallback to OCR for scanned documents
• 🌐 Web Interface: Clean Gradio-based chat interface
• ⚡ MCP Integration: Leverages Model Context Protocol for seamless tool integration
• 🔧 Easy Setup: Simple configuration and deployment

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Gradio UI     │───▶│   Client Agent   │───▶│   Claude AI     │
│  (Chat Interface)│    │  (MCP Client)    │    │   (Anthropic)   │
└─────────────────┘    └──────────────────┘    └─────────────────┘
                                │
                                ▼
                       ┌──────────────────┐
                       │   MCP Server     │
                       │  (FastMCP)       │
                       └──────────────────┘
                                │
                                ▼
                       ┌──────────────────┐
                       │   docs/ folder   │
                       │  (.md, .txt,     │
                       │   .rst files)    │
                       └──────────────────┘

• Python 3.10 or higher
• An Anthropic API key
• UV package manager (recommended) or pip

git clone <your-repo-url>
cd docs-navigator

Using UV (recommended):

uv sync

Or using pip:

pip install -r requirements.txt

Create a .env file:

echo "ANTHROPIC_API_KEY=your_api_key_here" > .env

Place your documentation files in the docs/ directory:

docs/
├── overview.md
├── setup.md
├── troubleshooting.md
└── your-other-docs.txt

# Using UV
uv run app_gradio.py

# Or directly with Python
python app_gradio.py

The app will be available at http://127.0.0.1:7860

Once the app is running, you can ask questions like:

• "How do I set up the authentication?"
• "What are the troubleshooting steps for connection issues?"
• "Where can I find information about API endpoints?"
• "Summarize the main features mentioned in the docs"

The AI will search through your documentation and provide contextual answers with references to the source files.

docs-navigator/
├── app_gradio.py          # Gradio web interface
├── client_agent.py        # MCP client and Claude integration
├── server_docs.py         # MCP server with doc tools
├── docs/                  # Your documentation files
├── tests/                 # Test scripts
│   ├── test_mcp.py       # Test MCP server functionality
│   ├── test_anthropic.py # Test Claude API connection
│   └── test_complete.py  # End-to-end functionality test
├── .env                   # Environment variables
├── pyproject.toml         # Project configuration
└── requirements.txt       # Python dependencies

The server exposes these tools to the AI:

1. list_docs(): Get a list of all available documentation files
2. search_docs(query, max_results): Search for specific content across all docs

Run the test suite:

# Test MCP server functionality
python test_mcp.py

# Test Claude API connection
python test_anthropic.py

# Test complete end-to-end functionality
python test_complete.py

• Markdown: .md
• Text: .txt
• reStructuredText: .rst
• PDF Documents: .pdf (text extraction + OCR for image-based content)
• Image Files: .png, .jpg, .jpeg, .tiff, .tif, .bmp, .gif (OCR text extraction)

For image-based PDF and image file processing, install Tesseract OCR:

Windows:

# Download and install from: https://github.com/UB-Mannheim/tesseract/wiki
# Or install via chocolatey:
choco install tesseract

macOS:

brew install tesseract

Ubuntu/Debian:

sudo apt-get install tesseract-ocr
sudo apt-get install poppler-utils  # For PDF image conversion

Additional Language Packs (optional):

• Windows: Select during Tesseract installation
• macOS: brew install tesseract-lang
• Ubuntu: sudo apt-get install tesseract-ocr-[language_code]

The app currently uses claude-3-haiku-20240307. To change the model, edit the model name in client_agent.py:

model="claude-3-haiku-20240307"  # Change to your preferred model

1. "Model not found" error: Your API key may not have access to the specified Claude model. The app will automatically test available models.

2. MCP connection issues: Ensure the server_docs.py script is executable and in the correct location.

3. No documents found: Make sure your documentation files are in the docs/ folder with supported extensions.

4. Gradio interface not loading: Check that port 7860 is available or modify the port in app_gradio.py.

Enable verbose logging by modifying the logging level in the respective files.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

This project is licensed under the MIT License - see the LICENSE file for details.

• Built with FastMCP for Model Context Protocol integration
• Powered by Anthropic Claude for AI responses
• UI created with Gradio
• Package management with UV

For more detailed instructions, see the Getting Started Guide.