PDF Tools MCP

A Model Context Protocol (MCP) server for PDF manipulation tools using PyMuPDF with OCR capabilities. Convert PDFs to PNG pages, extract embedded images, extract text with OCR support, and get PDF metadata.

Features

• pdf_to_images: Convert PDF pages to PNG images
• extract_pdf_images: Extract embedded images from PDF files
• extract_pdf_text: Extract text from PDF pages
• extract_pdf_text_with_ocr: Extract text using OCR (Tesseract) - ideal for scanned PDFs or protected documents
• get_pdf_summary: Get PDF metadata and summary information

Version 0.3.0 Changes

• Added OCR functionality using Tesseract for text extraction from scanned/protected PDFs
• Enhanced image processing with OpenCV for better OCR accuracy
• Configurable page range processing (start_page, pages_to_read)
• Confidence scoring for OCR results
• Support for image enhancement and noise reduction

Prerequisites

For OCR functionality, you need Tesseract OCR installed:

Windows: Download and install from: https://github.com/UB-Mannheim/tesseract/wiki Default installation path: C:\Program Files\Tesseract-OCR\tesseract.exe

Linux:

sudo apt-get install tesseract-ocr

macOS:

brew install tesseract

Usage examples (local):

Install dependencies:

python -m venv .venv
source .venv/Scripts/activate
pip install -r requirements.txt

Convert PDF to pages:

python run_mcp.py pdf2png --read_file_path "/abs/path/to/file.pdf" --write_folder_path "/abs/out"

Copy cover and closing into a theme:

python copy_cover_closing.py "/abs/out/pages" "C:/Users/USER/MCPServers/haia-slidev-builder-mcp/themes/GreenCoin-theme/assets"

Usage via npm (after publish):

Run without install:

npx pdf-tools-mcp --help
npx pdf-tools-mcp pdf2png --read_file_path "/abs/file.pdf" --write_folder_path "/abs/out"

Use as an MCP tool (example client config):

{
	"mcpServers": {
		"pdf-tools": {
			"command": "node",
			"args": ["pdf-tools-mcp", "--"],
			"env": {
				// Ensure Python is available on PATH; optional to set PYTHON explicitly
				// "PYTHON": "C:/Users/USER/.virtualenvs/pdf-tools/Scripts/python.exe"
			}
		}
	}
}