Migueel0/mcp-pdf-reader

mcp-pdf-reader

MCP server that extracts text from PDF files using pypdf.
Optionally, if Tesseract OCR is installed, the server can also extract text from images embedded inside PDFs.

The server exposes a single MCP tool: read_pdf(file_path), which returns a JSON-serializable dictionary containing the original file path and the extracted text.

Project Structure

• server/ --- MCP tool definitions used by the server
• tools/ --- Utility modules; pdf_reader.py contains PDF extraction logic
• main.py --- MCP server launcher
• requirements.txt --- Runtime dependencies

Quick Summary

Tool: read_pdf(file_path)

Returns:

{
  "file_path": "path/to/file.pdf",
  "extracted_text": "Sample extracted text..."
}

Requirements

• Python ≥ 3.13 (recommended)
• virtualenv (recommended)
• Tesseract OCR (optional, required only for OCR on images inside PDFs)

Installing Dependencies (Windows)

1. Create and activate a virtual environment:

python -m venv venv
.venv\Scripts\activate

2. Verify your Python version:

python --version

3. Install dependencies:

python -m pip install --upgrade pip
pip install -r requirements.txt

Installing Dependencies (Linux)

1. Create and activate a virtual environment:

python3 -m venv venv
source venv/bin/activate

2. Verify your Python version:

python3 --version

3. Install dependencies:

python3 -m pip install --upgrade pip
pip install -r requirements.txt

Installing Tesseract (Optional)

Follow the official installation instructions:
https://tesseract-ocr.github.io/tessdoc/Installation.html

You may need to manually add tesseract to your system PATH.

More details available in the Tesseract User Manual:
https://github.com/tesseract-ocr/tessdoc

Configuring Environment Variables

Rename env.example to .env and set your Tesseract binary path.

Example:

# Windows
TESSERACT_CMD=C:\Program Files\Tesseract-OCR\tesseract.exe

# Linux
TESSERACT_CMD=/usr/bin/tesseract

Running the Server

Launch the MCP server using:

python main.py

If everything is configured correctly, the server will start and expose the read_pdf tool.

MCP Usage Example

Call the tool from your MCP client:

Tool exposed:

read_pdf(file_path)

Example response:

{
  "file_path": "sample/path/sample_file.pdf",
  "extracted_text": "Sample text
..."
}

MCP Launcher Configuration

Windows

{
  "mcpServers": {
    "pdf-reader": {
      "command": "PATH/TO/VENV/Scripts/python.exe",
      "args": [
        "PATH/TO/REPO/mcp-pdf-reader/main.py"
      ]
    }
  }
}

Linux

{
  "mcpServers": {
    "pdf-reader": {
      "command": "PATH/TO/VENV/bin/python",
      "args": [
        "PATH/TO/REPO/mcp-pdf-reader/main.py"
      ]
    }
  }
}

Notes

This project is an early test version. Future updates will extend available tools and improve installation and deployment workflows.