daevosDev/markitdown-mcp-ocr

MarkItDown-MCP with OCR Support

A powerful MCP (Model Context Protocol) server that provides document conversion capabilities with OCR support using Tesseract OCR.

Features

• Document Conversion: Convert various document formats to markdown
• OCR Support: Extract text from images using Tesseract OCR
• Multiple Languages: Support for English and French OCR
• Docker Support: Ready-to-use Docker containers
• SSE Mode: Server-Sent Events support for real-time updates

Quick Start

Using Docker (Recommended)

Basic Docker Image

# Build the basic image
docker build -t markitdown-mcp .

# Run the container
docker run -p 3001:3001 --name markitdown-mcp markitdown-mcp

OCR-Enabled Docker Image

# Build the OCR-enabled image
docker build -f Dockerfile.ocr -t markitdown-mcp-ocr .

# Run the OCR container with volume mounting
docker run -p 3002:3001 -v $(pwd)/TEST:/app/test_files --name markitdown-mcp-ocr markitdown-mcp-ocr

Alternative OCR Configuration

# Build the alternative OCR image
docker build -f Dockerfile.test-ocr-fixed -t markitdown-mcp-ocr-fixed .

# Run with SSE on port 8000
docker run -p 8000:8000 -v $(pwd)/TEST:/app/test_files --name markitdown-mcp-sse markitdown-mcp-ocr-fixed

Manual Installation

Prerequisites

• Python 3.11+
• Tesseract OCR (for OCR support)

Install Dependencies

pip install -r requirements.txt

Run the Server

# Standard HTTP mode
markitdown-mcp --http --host 0.0.0.0 --port 3001

# SSE mode
python server.py

Usage

API Endpoints

• HTTP Mode: http://localhost:3001/mcp
• SSE Mode: http://localhost:8000

Testing OCR

1. Place test images in the TEST/ directory
2. Use the provided test scripts:

   python test-ocr-fix.py
   

Docker Compose

Use the provided docker-compose.yml for easy deployment:

docker-compose up -d

File Structure

markitdown-mcp/
├── Dockerfile              # Basic Docker configuration
├── Dockerfile.ocr          # OCR-enabled Docker configuration
├── Dockerfile.test-ocr-fixed # Alternative OCR configuration
├── docker-compose.yml      # Docker Compose setup
├── requirements.txt        # Python dependencies
├── server.py              # SSE server script
├── test-ocr-fix.py        # OCR testing script
├── TEST/                  # Test files directory
│   ├── demo-ocr-test.png
│   ├── test-ocr-fr.png
│   └── ...
└── README.md              # This file

Configuration

Environment Variables

• HOST: Server host (default: 0.0.0.0)
• PORT: Server port (default: 3001 for HTTP, 8000 for SSE)

OCR Languages

The OCR configuration supports:

• English (eng)
• French (fra)

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test with Docker
5. Submit a pull request

License

This project is licensed under the MIT License.