home / mcp / doc converter mcp server

DOC Converter MCP Server

Converts DOCX to Markdown, extracts images, and analyzes image data via MCP over stdio.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "kittolau-mcp": {
      "command": "node",
      "args": [
        "src/server.js"
      ]
    }
  }
}

You are using a compact MCP server that converts DOCX documents to Markdown while extracting embedded images and offers image analysis tools. It runs using the MCP protocol over stdio, making it easy to connect from editors like VS Code Copilot or other MCP clients for seamless document processing and automation tasks.

How to use

You connect to this server from an MCP client to perform document conversion and image-related tasks. Once connected, you can run conversion to Markdown, extract images from documents, read image metadata, and generate simple HTML views for image sets. Use it as part of your automated workflows or editor-assisted tooling to turn DOCX content into portable Markdown while preserving formatting and structure.

How to install

Prerequisites you need before installing this server are: Node.js 14 or higher. You should also have npm available to install dependencies and to run start scripts.

# Install the MCP server globally
npm install -g mcp-document-converter

# Start the server using the programmatic entry point
node src/server.js

# Or start via the global command if available after installation
mcp-document-converter

Configuration and starting the server

Choose the starting method that fits your workflow. If you prefer running directly from source, use the programmatic entry point. If you want a global command, install the package globally and start with the provided binary.

# Start via programmatic entry point
node src/server.js

# Start via the installed global command (if present after installation)
mcp-document-converter

Notes and tips

The server communicates using JSON-RPC over stdio. You will send requests on stdin and receive responses on stdout. This enables seamless integration with editors and automation pipelines.

Troubleshooting

If the server fails to start, verify that Node.js 14+ is installed and that npm dependencies are available. Check that you are using the correct start command for your setup (programmatic or global). If you encounter permission issues with a global install, consider using a local installation and invoking node src/server.js directly.

Available tools

convert_docx_to_markdown

Converts DOCX documents to Markdown while preserving text formatting, headers, lists, and blockquotes, and optionally extracts embedded images into a dedicated folder with relative links.

read_image_info

Reads metadata and dimensions of images to help you understand the content and size implications in your documents.

analyze_images_directory

Analyzes all images within a folder to provide a quick overview of image kinds, counts, and potential optimizations.

create_image_viewer

Generates an HTML viewer that lets you browse extracted images within a simple page for quick review.

open_browser

Opens a browser window to a specified URL or performs a search when used through MCP tooling, enabling quick web access from your automation flow.

browser_input_text

Automates typing text into web forms by navigating to a URL, selecting the input field, entering text, and optionally submitting the form with an option to take a screenshot.