PDFtotext MCP Server

A reliable Model Context Protocol (MCP) server for PDF text extraction using the proven pdftotext utility from poppler-utils.

🚀 Why This Server?

Unlike other PDF MCP servers that suffer from logging interference, complex dependencies, and reliability issues, pdftotext-mcp is:

• ✅ Actually works - Clean JSON-RPC communication without stdout pollution
• ✅ Reliable - Built on mature pdftotext from poppler-utils (used by millions)
• ✅ Lightweight - Minimal dependencies, maximum compatibility
• ✅ Production tested - Successfully tested with Claude Desktop and other MCP clients
• ✅ Feature complete - Page-specific extraction, layout preservation, encoding options
• ✅ Error handling - Comprehensive validation and helpful error messages

📋 Features

• 📄 Extract text from entire PDF documents or specific pages
• 🎨 Preserve original layout formatting (optional)
• 🔤 Multiple text encoding support (UTF-8, Latin1, ASCII)
• 📊 Comprehensive metadata in responses (word count, file info, etc.)
• 🛡️ File validation and security checks
• ⚡ Fast processing with configurable timeouts
• 🔍 Detailed error reporting with troubleshooting hints

🔧 Prerequisites

You must have pdftotext installed on your system:

Ubuntu/Debian

sudo apt update
sudo apt install poppler-utils

macOS

brew install poppler

Windows

# Using Chocolatey
choco install poppler

# Using Scoop
scoop install poppler

Verify Installation

pdftotext -v

📦 Installation

Option 1: Global Installation (Recommended)

npm install -g pdftotext-mcp

Option 2: Use with npx (No Installation)

npx pdftotext-mcp

Option 3: Local Development

git clone https://github.com/jpwebb/pdftotext-mcp.git
cd pdftotext-mcp
npm install
npm start

⚙️ Configuration

Add to your MCP client configuration:

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "pdftotext": {
      "command": "pdftotext-mcp"
    }
  }
}

Or with npx:

{
  "mcpServers": {
    "pdftotext": {
      "command": "npx",
      "args": ["pdftotext-mcp"]
    }
  }
}

Other MCP Clients

The server works with any MCP-compatible client. Use pdftotext-mcp as the command.

🎯 Usage

The server provides a single, powerful tool: read_pdf_text

Basic Usage

Extract entire document

{
  "tool": "read_pdf_text",
  "arguments": {
    "path": "./document.pdf"
  }
}

Extract specific page

{
  "tool": "read_pdf_text",
  "arguments": {
    "path": "./document.pdf",
    "page": 2
  }
}

Preserve layout formatting

{
  "tool": "read_pdf_text",
  "arguments": {
    "path": "./document.pdf",
    "layout": true
  }
}

Custom encoding

{
  "tool": "read_pdf_text",
  "arguments": {
    "path": "./document.pdf",
    "encoding": "Latin1"
  }
}

Response Format

Success Response

{
  "success": true,
  "file": "document.pdf",
  "path": "/absolute/path/to/document.pdf",
  "extractedText": "Full text content...",
  "pageSpecific": "all",
  "layoutPreserved": false,
  "encoding": "UTF-8",
  "fileSize": 1048576,
  "lastModified": "2024-01-15T10:30:00.000Z",
  "extractedAt": "2024-01-15T10:35:00.000Z",
  "textLength": 5234,
  "wordCount": 892
}

Error Response

{
  "success": false,
  "error": "File not found: ./nonexistent.pdf",
  "errorType": "FILE_NOT_FOUND",
  "file": "./nonexistent.pdf",
  "timestamp": "2024-01-15T10:35:00.000Z"
}

📚 API Reference

Tool: read_pdf_text

Extracts text content from PDF files using pdftotext.

Parameters

Supported Encodings

• UTF-8 (default)
• Latin1
• ASCII

Error Types

• FILE_NOT_FOUND - PDF file doesn't exist
• PERMISSION_DENIED - Cannot read the file
• INVALID_PDF - File is not a valid PDF
• PDFTOTEXT_ERROR - pdftotext utility error
• UNKNOWN_ERROR - Unexpected error

🔧 Troubleshooting

"pdftotext is not available"

Solution: Install poppler-utils (see Prerequisites)

"File not found"

Solutions:

• Use absolute paths: /home/user/document.pdf
• Check file exists: ls -la /path/to/file.pdf
• Verify MCP server working directory

"Permission denied"

Solutions:

• Check file permissions: chmod 644 document.pdf
• Ensure directory is readable: chmod 755 /path/to/directory/

"File is not a valid PDF"

Solutions:

• Verify file is actually a PDF: file document.pdf
• Check for file corruption
• Try with a different PDF file

MCP Connection Issues

Solutions:

• Restart your MCP client completely
• Check configuration syntax in config file
• Verify pdftotext-mcp is accessible in PATH
• Check MCP client logs for detailed errors

🧪 Testing

# Run tests
npm test

# Run tests with watch mode
npm run test:watch

# Run linter
npm run lint

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Development Setup

git clone https://github.com/jpwebb/pdftotext-mcp.git
cd pdftotext-mcp
npm install

Running Locally

npm start

Code Style

This project uses ESLint. Run npm run lint to check code style.

📄 License

MIT - see LICENSE file for details.

🙏 Acknowledgments

• Built for the Model Context Protocol ecosystem
• Uses poppler-utils pdftotext utility
• Inspired by the need for reliable PDF processing in MCP environments

🔗 Related

• Model Context Protocol Documentation
• Claude Desktop MCP Configuration
• Poppler Utils Documentation

---

Made for the MCP community