Skip to main content

MCP server for bidirectional Markdown ↔ Word (.docx) conversion

Project description

MD-DOCX Converter

A Python tool for bidirectional conversion between Markdown (.md) and Microsoft Word (.docx). Designed to make it easy to move content between Word documents and AI tools like Claude, ChatGPT, and GitHub Copilot.

What it does

  • Converts .md.docx with correct heading hierarchy (Title, Heading 1–9)
  • Converts .docx.md as clean GitHub Flavored Markdown (GFM)
  • Runs from a simple desktop shortcut — no command line knowledge needed
  • Handles headings, bold/italic/strikethrough, lists, task lists, tables, blockquotes, code blocks, images, and hyperlinks

See MarkdownSyntax.md for the full element mapping and notes on what is preserved, approximated, or dropped.

Requirements

  • Windows 10/11
  • Python 3.11+
  • The following Python packages (installed via pip):
pip install markdown-it-py python-docx

Setup

1. Clone the repository

git clone https://github.com/cjwpenner/md-docx-converter.git
cd md-docx-converter

2. Install dependencies

pip install markdown-it-py python-docx

3. Create the desktop shortcut

pip install pywin32
python create_shortcut.py

This creates an MD-DOCX Converter shortcut on your Windows desktop. pywin32 is only needed to create the shortcut — it is not required to run the converter itself.

4. Run the converter

Double-click MD-DOCX Converter on your desktop. A console window opens and prompts:

MD ↔ DOCX Converter
--------------------
Enter file path:

Paste or type the full path to your .md or .docx file and press Enter. The converted file is saved in the same directory with the extension swapped.

You can also run directly from the command line:

python md_docx_converter/converter.py

Conversion notes

Heading hierarchy

The heading level mapping is context-dependent:

  • MD → DOCX: If there is exactly one # in the document, it becomes a Word Title. All other headings shift down by one level. If there are multiple # headings, they all become Heading 1 with no Title.
  • DOCX → MD: If the document has a Title style, it becomes #. All headings shift up accordingly. If there is no Title, Heading 1 becomes #.

Lossy elements

Word formatting that has no Markdown equivalent is approximated as bold:

Word formatting Markdown output
Underline **bold**
Highlight **bold**
Small caps **bold**
Font colour Stripped (text kept)

Images

  • DOCX → MD: Embedded images are extracted to a {filename}_images/ folder next to the output .md file.
  • MD → DOCX: Images referenced by relative path are re-embedded. Missing images become [image not found: path].

MCP server (Claude / AI integration)

This tool is also available as an MCP server, letting Claude and other AI assistants read and write Word documents directly.

Install

pip install mcp-md-docx

Configure Claude Desktop

Add to %APPDATA%\Claude\claude_desktop_config.json:

{
  "mcpServers": {
    "md-docx": {
      "command": "python",
      "args": ["-m", "mcp_md_docx"]
    }
  }
}

Configure Claude Code

claude mcp add md-docx -- python -m mcp_md_docx

Tools exposed

Tool What it does
read_docx Read a .docx file — returns full Markdown text to the AI
write_docx Create a .docx from Markdown text the AI has written
convert_md_file_to_docx Convert a .md file on disk to .docx
convert_docx_file_to_md Convert a .docx file on disk to .md

Once configured, you can say things like:

  • "Read report.docx and summarise it"
  • "Turn this into a Word document and save it to my Desktop"
  • "Convert all the bullet points in notes.docx into a table"

Project structure

md_docx_converter/
├── converter.py       # CLI entry point
├── md_to_docx.py      # Markdown → Word conversion
├── docx_to_md.py      # Word → Markdown conversion
├── heading_mapper.py  # Heading hierarchy pre-scan logic
├── image_handler.py   # Image extraction and embedding
└── launch.pyw         # Desktop shortcut launcher
mcp_md_docx/
├── server.py          # MCP server (four tools)
└── __main__.py        # Entry point for python -m mcp_md_docx
create_shortcut.py     # One-time shortcut setup script
pyproject.toml         # PyPI packaging config

License

This project is licensed under the GNU General Public License v3.0 (GPLv3). You are free to use, modify, and distribute this software, provided that any derivative works are also distributed under the same licence.

See LICENSE for the full licence text.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mcp_md_docx-0.1.0.tar.gz (251.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mcp_md_docx-0.1.0-py3-none-any.whl (25.7 kB view details)

Uploaded Python 3

File details

Details for the file mcp_md_docx-0.1.0.tar.gz.

File metadata

  • Download URL: mcp_md_docx-0.1.0.tar.gz
  • Upload date:
  • Size: 251.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.12

File hashes

Hashes for mcp_md_docx-0.1.0.tar.gz
Algorithm Hash digest
SHA256 99e4cf2c37bbd32354b3bc149a3db74928cb926b6cea1a124ea6c150110ec19d
MD5 c980d52692263f86945ff842294a2ca2
BLAKE2b-256 9d4b1cc44ae837bbbb7722333667fecfa991933c64bf6ec623e5d0f6c1eeb433

See more details on using hashes here.

File details

Details for the file mcp_md_docx-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: mcp_md_docx-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 25.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.12

File hashes

Hashes for mcp_md_docx-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 a2801844ae4eb556986fa8171b93aec96594f14a2e979dce32a612260a5bff94
MD5 a53b16b2a35d7924fee26be708f590f2
BLAKE2b-256 80831cf8798fa294087a26357d2c613b0f56d5017fb5a4957958b981dcbdd28a

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page