word-mcp-live
Enables AI assistants to edit Microsoft Word documents live while they are open, with full support for tracked changes, comments, formatting, and 124 tools.
README
<div align="center">
word-mcp-live
The only MCP server that edits Word documents while they're open
Live editing · Tracked changes · Per-action undo · 124 tools · Cross-platform
</div>
word-mcp-live gives any AI assistant that supports MCP full control of Microsoft Word. Open a document, tell the AI what you need, and watch it happen — formatting, tracked changes, comments, and all. Changes appear live in your open document.
<table> <tr> <td width="50%">
Without word-mcp-live
- AI can discuss your document but can't touch it
- You copy-paste between AI and Word, losing formatting
- Track changes? You do those manually after the fact
- Every edit means save → close → process → reopen
</td> <td width="50%">
With word-mcp-live
- "Add a tracked change replacing ABC Corp with XYZ Ltd" — done
- Changes appear live in your open Word document
- Every AI edit is one Ctrl+Z away
- Real tracked changes with your name, not XML hacks
</td> </tr> </table>
See it in action
https://github.com/user-attachments/assets/fbb09af4-1e25-4e49-94d0-45b363278810
What Sets This Apart
- Live editing — Edit documents while they're open in Word. No save-close-reopen cycle.
- Full undo — Every AI action is a single Ctrl+Z. Made a mistake? Just undo it.
- Native tracked changes — Real Word revisions with your name, not XML hacks.
- Threaded comments — Add, reply, resolve, and delete comments like a human reviewer.
- Layout diagnostics — Detects formatting problems before they become print disasters.
- Equations & cross-references — Insert math formulas and auto-updating references.
- 124 tools — The most comprehensive Word MCP server available.
Quick Start
pip install word-mcp-live
Or install from source:
git clone https://github.com/ykarapazar/word-mcp-live.git
cd word-mcp-live
pip install -e .
Client Installation
<details open> <summary><b>Claude Desktop</b></summary>
Add to your claude_desktop_config.json:
{
"mcpServers": {
"word": {
"command": "uvx",
"args": ["word-mcp-live"],
"env": {
"MCP_AUTHOR": "Your Name",
"MCP_AUTHOR_INITIALS": "YN"
}
}
}
}
</details>
<details> <summary><b>Claude Code</b></summary>
Add to your .mcp.json:
{
"mcpServers": {
"word": {
"command": "uvx",
"args": ["word-mcp-live"],
"env": {
"MCP_AUTHOR": "Your Name",
"MCP_AUTHOR_INITIALS": "YN"
}
}
}
}
</details>
<details> <summary><b>Cursor</b></summary>
One-click: Click the install button at the top of this page.
Manual: Add to ~/.cursor/mcp.json:
{
"mcpServers": {
"word": {
"command": "uvx",
"args": ["word-mcp-live"],
"env": {
"MCP_AUTHOR": "Your Name",
"MCP_AUTHOR_INITIALS": "YN"
}
}
}
}
</details>
<details> <summary><b>VS Code / Copilot</b></summary>
One-click: Install in VS Code
Manual: Add to your VS Code settings.json:
{
"mcp": {
"servers": {
"word": {
"command": "uvx",
"args": ["word-mcp-live"],
"env": {
"MCP_AUTHOR": "Your Name",
"MCP_AUTHOR_INITIALS": "YN"
}
}
}
}
}
</details>
<details> <summary><b>Windsurf</b></summary>
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"word": {
"command": "uvx",
"args": ["word-mcp-live"],
"env": {
"MCP_AUTHOR": "Your Name",
"MCP_AUTHOR_INITIALS": "YN"
}
}
}
}
</details>
<details> <summary><b>Docker</b></summary>
{
"mcpServers": {
"word": {
"command": "docker",
"args": ["run", "-i", "--rm", "ghcr.io/ykarapazar/word-mcp-live"],
"env": {
"MCP_AUTHOR": "Your Name",
"MCP_AUTHOR_INITIALS": "YN"
}
}
}
}
Note: Docker mode supports cross-platform tools only. Live editing requires a native Windows install.
</details>
MCP_AUTHORsets your name on tracked changes and comments (default:"Author").MCP_AUTHOR_INITIALSsets comment initials.
Two Modes
| Works everywhere | Live editing (Word open) | |
|---|---|---|
| What it does | Create and edit saved .docx files | Edit documents live while you work in Word |
| Platform | Windows, macOS, Linux | Windows (COM) and macOS (JXA) |
| Undo | File-level saves | Per-action Ctrl+Z (Windows); per-operation undo (macOS) |
| Best for | Batch processing, document generation | Interactive editing, formatting, review |
Both modes work together. The AI picks the right one for the task.
macOS Live Editing (New in v1.5.0)
Live tools now work on macOS via JavaScript for Automation (JXA). Same tool names, same parameters — the server detects your platform and uses the right automation backend.
| Feature | Windows | macOS |
|---|---|---|
| Text read/write/find/replace | COM | JXA |
| Formatting (bold, font, style) | COM | JXA |
| Track changes & revisions | COM | JXA |
| Comments (add, delete, list) | COM | JXA |
| Tables (read, write, add rows) | COM | JXA |
| Page layout, headers, bookmarks | COM | JXA |
| Equations, cross-references | COM | JXA |
| Threaded comment replies | COM | Not available |
| Comment resolve/unresolve | COM | Not available |
| Undo history inspection | COM | Not available |
| Watermarks | COM | Not available |
Configuration
| Variable | Default | Description |
|---|---|---|
MCP_AUTHOR |
"Author" |
Author name for tracked changes and comments |
MCP_AUTHOR_INITIALS |
"" |
Author initials for comments |
MCP_TRANSPORT |
stdio |
Transport type: stdio, sse, or streamable-http |
MCP_HOST |
0.0.0.0 |
Host to bind (for SSE/HTTP transports) |
MCP_PORT |
8000 |
Port to bind (for SSE/HTTP transports) |
For remote deployment, see RENDER_DEPLOYMENT.md.
Example Prompts
Just tell the AI what you want in plain language:
"Draft a contract with tracked changes so my colleague can review"
"Format all headings as Cambria 13pt bold and add automatic numbering"
"Add a comment on paragraph 3 asking about the deadline"
"Find every mention of 'ABC Corp' and replace with 'XYZ Ltd' as a tracked change"
"Set the page to A4 landscape with 2cm margins"
"Insert a table of contents based on the document headings"
"Add page numbers in the footer and our company name in the header"
"Insert a cross-reference to Heading 2 in paragraph 5"
Usage Examples
Example 1: Read a document (cross-platform)
Tool call: get_document_text
{ "filename": "C:/Documents/report.docx" }
Expected output:
{
"status": "success",
"paragraphs": [
{"index": 0, "text": "Quarterly Report", "style": "Heading 1"},
{"index": 1, "text": "Revenue increased by 15% compared to Q3.", "style": "Normal"},
{"index": 2, "text": "Key Metrics", "style": "Heading 2"}
],
"total_paragraphs": 3
}
Example 2: Live editing with tracked changes (Windows)
Tool call: word_live_replace_text
{
"filename": "report.docx",
"find_text": "ABC Corporation",
"replace_text": "XYZ Ltd",
"match_case": true,
"replace_all": true,
"track_changes": true
}
Expected output:
{
"status": "success",
"replacements": 4,
"message": "Replaced 4 occurrences (tracked changes enabled)"
}
The replacements appear as tracked changes in Word with strikethrough on "ABC Corporation" and underline on "XYZ Ltd".
Example 3: Add a comment anchored to text (cross-platform)
Tool call: add_comment
{
"filename": "C:/Documents/contract.docx",
"target_text": "payment within 30 days",
"comment_text": "Should we extend this to 45 days?",
"author": "Jane Smith"
}
Expected output:
{
"status": "success",
"message": "Comment added by Jane Smith on 'payment within 30 days'"
}
The comment appears in Word's Review panel, anchored to the specified text.
Tool Reference
124 tools across two modes — see the complete tool reference for details.
| Category | Count |
|---|---|
| Cross-platform (python-docx) | 80 |
| Windows Live (COM automation) | 44 |
| macOS Live (JXA automation) | 40 (of the 44 live tools) |
Requirements
- Python 3.11+
python-docx,fastmcp,msoffcrypto-tool(installed automatically)- Windows Live tools: Windows 10/11 + Microsoft Word +
pywin32(installed automatically) - macOS Live tools: macOS + Microsoft Word for Mac (uses built-in JXA — no extra dependencies)
The cross-platform tools work without Word installed — only python-docx is needed.
Contributing
See CONTRIBUTING.md for development setup, code style, and how to add new tools.
Found a bug? Open an issue. Have an idea? Request a feature.
Acknowledgments
Built on top of GongRzhe/Office-Word-MCP-Server by GongRzhe (MIT License).
Additional libraries: python-docx · FastMCP · pywin32
Privacy
This server runs entirely on your local machine. No data is collected, transmitted, or stored. See the full Privacy Policy.
Support
- Bug reports: Open an issue
- Feature requests: Request a feature
- Discussions: GitHub Discussions
License
MIT License — see LICENSE for details.
Star History
<a href="https://star-history.com/#ykarapazar/word-mcp-live&Date"> <picture> <source media="(prefers-color-scheme: dark)" srcset="https://api.star-history.com/svg?repos=ykarapazar/word-mcp-live&type=Date&theme=dark" /> <source media="(prefers-color-scheme: light)" srcset="https://api.star-history.com/svg?repos=ykarapazar/word-mcp-live&type=Date" /> <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=ykarapazar/word-mcp-live&type=Date" /> </picture> </a>
<!-- mcp-name: io.github.ykarapazar/word-mcp-live -->
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.