Word MCP Server

What is Word MCP Server?

Word MCP Server is a Python application that creates and edits Microsoft Word (.docx) documents through an API. It uses FastMCP to expose tools for interacting with Word documents, primarily for integration with large language models (LLMs).

How to use Word MCP Server?

Install dependencies with uv pip install . in a Python 3.12+ virtual environment. Configure the server by editing a JSON file with the Python interpreter path and server script path, then launch it through an LLM client. The server exposes functions such as create_new_document(), add_heading(), add_paragraph(), add_run_to_paragraph(), add_picture(), add_table(), and save_document().

Key features of Word MCP Server

• Create and open Word documents (.docx)
• Add and format text with headings, paragraphs, and runs
• Insert images from file paths or numpy arrays
• Create tables with custom styles
• Manage resources and prompts stored in directories
• Integrate with LLMs via the MCP protocol

Use cases of Word MCP Server

• Automatically generate formatted reports from LLM conversations
• Create Word documents programmatically with structured content
• Edit existing documents by adding images, tables, and styled text
• Build document templates populated by AI prompts

FAQ from Word MCP Server

What does Word MCP Server do?

It provides tools for creating, editing, and formatting Microsoft Word documents (.docx) programmatically, with commands that can be triggered by large language models through the MCP protocol.

What are the dependencies and runtime requirements?

Python 3.12 or higher is required. Libraries include python-docx, opencv-python (cv2), numpy, and FastMCP. All dependencies are installed via uv pip install ..

How do I configure Word MCP Server with an LLM?

You create a JSON configuration specifying the Python interpreter path and the server script path under mcpServers.word-mcp-server. Once launched, the LLM can call the server's tools using natural language.

What image formats and colors are supported?

Images can be added from file paths or numpy arrays. For text formatting, supported color values include black, blue, green, red, yellow, pink, teal, violet, gray25, gray50, and several dark variants. Highlight values are similar.

Where does data live and what are the limits?

Documents, resources, and prompts are stored locally in files and directories (resources/ and prompts/). The server only works with .docx files; no authentication or remote storage is mentioned.