pdf-reader-mcp

PDF Reader MCP Server

A powerful Model Context Protocol (MCP) server that provides comprehensive PDF processing capabilities, including text extraction, OCR support, and network diagram detection. Designed to work seamlessly with Amazon Q Developer CLI and other MCP-compatible systems.

🚀 Features

• Text Extraction: Extract text content from PDF files with high accuracy
• PDF Analysis: Get comprehensive metadata (pages, document info, encryption status)
• Page-Specific Processing: Extract text from specific pages
• Multi-Language OCR: Support for Thai and English text recognition
• Smart Processing: Automatically chooses between OCR and direct text extraction
• Markdown Conversion: Convert PDF content to clean markdown format
• Document Analysis: Determine if PDFs are scanned images or searchable text
• Network Diagram Detection: Advanced capability to detect and extract network diagrams
• MCP Integration: Seamless integration with Amazon Q Developer CLI

📦 Installation

Using uvx (Recommended)

The easiest way to use this MCP server is with uvx:

# Run directly without installation
uvx pdf-reader-mcp # your mcp client can now connect using both sse and stdio transport

# Or install globally
uvx install pdf-reader-mcp

Using pip

pip install pdf-reader-mcp

From Source

git clone https://github.com/zixma13/pdf-reader-mcp.git
cd pdf-reader-mcp
pip install -e .

📋 Prerequisites

• Install Tesseract OCR (required for OCR functionality):

  # For macOS
  brew install tesseract
  brew install tesseract-lang  # For language support including Thai
  
  # For Ubuntu/Debian
  # sudo apt-get install tesseract-ocr
  # sudo apt-get install tesseract-ocr-tha  # For Thai language support
  

• Install Python dependencies:

  pip install -r requirements.txt
  

• Ensure you have the virtual environment activated:

  source .venv/bin/activate
  

• Test the server:

  mcp dev main.py
  

Configuration for Amazon Q Developer CLI

Add the following to your ~/.aws/amazonq/mcp.json file:

• in case you clone the git repository

{
  "mcpServers": {
    "pdf_reader": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/your/pdf_reader/pdf_reader",
        "run",
        "main.py"
      ]
    }
  }
}

• in case you use uvx

{
  "mcpServers": {
    "pdf_reader": {
      "command": "uvx",
      "timeout": 60000,
      "args": [
        "pdf-reader-mcp"
      ]
    }
  }
}

Usage Examples

Once configured, you can use the PDF reader tools in Amazon Q:

Basic PDF Processing

• To analyze a PDF and determine if it's a scanned image or searchable text:

  pdf_reader___analyze_pdf("/path/to/document.pdf")
  

• To intelligently extract content from a PDF (automatically choosing between OCR and text extraction):

  pdf_reader___smart_extract_pdf("/path/to/document.pdf")
  

• To intelligently convert a PDF to markdown:

  pdf_reader___smart_pdf_to_markdown("/path/to/document.pdf")
  

• To extract text from a PDF:

  pdf_reader___read_pdf("/path/to/document.pdf")
  

• To get metadata from a PDF:

  pdf_reader___get_pdf_metadata("/path/to/document.pdf")
  

• To extract text from a specific page (0-indexed):

  pdf_reader___extract_pdf_page("/path/to/document.pdf", 0)
  

• To extract text using OCR (supports Thai and English):

  pdf_reader___ocr_pdf("/path/to/document.pdf")
  

• To extract text from a specific page using OCR:

  pdf_reader___ocr_pdf_page("/path/to/document.pdf", 0)
  

• To convert PDF to markdown format:

  pdf_reader___pdf_to_markdown("/path/to/document.pdf")
  

• To convert PDF to markdown format using OCR:

  pdf_reader___pdf_to_markdown("/path/to/document.pdf", use_ocr=True)