OCR MCP Server

MCP (Model Context Protocol) server for OCR document conversion using XMO API.

Features

• Convert documents (PDF, Word, PowerPoint) and images to Markdown
• Supports multipart/form-data file upload
• API key authentication via XMO_APIKEY environment variable
• Support for advanced OCR options (LLM enhancement, page range, etc.)

Installation

Using npx (Recommended)

npx -y xmo-ocr-mcp-server

Install globally

npm install -g xmo-ocr-mcp-server

Install locally

npm install xmo-ocr-mcp-server

Usage

Set up API Key

You must set the XMO_APIKEY environment variable:

export XMO_APIKEY="your-api-key-here"

Run as MCP Server

The server runs on stdio and can be used with any MCP client:

XMO_APIKEY="your-api-key" npx -y xmo-ocr-mcp-server

Configure in Claude Desktop

Add to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "ocr": {
      "command": "npx",
      "args": ["-y", "xmo-ocr-mcp-server"],
      "env": {
        "XMO_APIKEY": "your-api-key-here"
      }
    }
  }
}

Available Tools

convert_document

Convert documents or images to Markdown using OCR.

Parameters:

• file_path (required): Absolute path to the local file
• use_llm (optional): Use LLM to enhance accuracy for tables, forms, math (default: false)
• output_format (optional): Output format - 'markdown', 'json', 'html', or 'chunks' (default: 'markdown')
• page_range (optional): Pages to process, e.g., '0,2-5' for pages 0,2,3,4,5
• max_pages (optional): Maximum number of pages to process
• force_ocr (optional): Force OCR on all PDF pages (default: false)

Example:

{
  "file_path": "/path/to/document.pdf",
  "use_llm": true,
  "page_range": "0-5"
}

Supported File Formats

• Documents: PDF, DOC, DOCX, PPT, PPTX
• Images: PNG, JPG, JPEG, WEBP

Development

Install dependencies

npm install

Run locally

export XMO_APIKEY="your-api-key"
node index.js

Test the package

export XMO_APIKEY="your-api-key"
npm test

Environment Variables

• XMO_APIKEY (required): Your XMO API key for authentication
• OCR_API_URL (optional): Custom OCR API endpoint (default: http://ocr.xm-opt.com/api/ocr/convert/sync)

Publishing to npm

Before publishing:

1. Ensure you're logged in to npm: npm login
2. Update version in package.json: npm version patch|minor|major
3. Publish to npm: npm publish --access public

Note: Make sure the package name xmo-ocr-mcp-server is available on npm before publishing.

License

MIT