A Model Context Protocol (MCP) server that allows AI agents to create, read, and edit Excel workbooks without requiring Microsoft Excel.
https://github.com/haris-musa/excel-mcp-serverStop wrestling with Excel licensing, installation headaches, and clunky automation scripts. This MCP server gives your AI agents full Excel manipulation capabilities using pure Python—no Microsoft Office required.
If you've ever tried automating Excel tasks, you know the pain: COM objects that break randomly, licensing restrictions across environments, and the constant "Excel is not installed" errors when deploying to servers. Meanwhile, your AI agents sit there helpless when you ask them to work with spreadsheet data.
This server solves both problems. Your AI can create workbooks, manipulate data, apply formatting, generate charts, and build pivot tables—all through conversational requests. No VBA, no brittle COM automation, no Excel installation whatsoever.
Complete Excel Operations: Read, write, format, create charts, build pivot tables, manage worksheets—everything you'd expect from full Excel automation, but through clean Python libraries (openpyxl, pandas, numpy, matplotlib).
Dual Transport Modes: Use stdio for local development with tools like Cursor, or SSE for remote server deployments. One server, flexible deployment options.
Zero Installation Friction: uvx excel-mcp-server stdio and you're running. No license keys, no registry modifications, no "please restart your computer" nonsense.
AI-Native Interface: Instead of writing complex automation scripts, just tell your AI what you want: "Create a sales dashboard with Q4 data and add a pivot table showing revenue by region."
Automated Reporting: Your AI can generate monthly reports by pulling data from APIs, creating formatted spreadsheets with charts and pivot tables, then saving them to shared drives—all from a simple request.
Data Pipeline Integration: Transform CSV exports into polished Excel deliverables with proper formatting, formulas, and visualizations as part of your automated workflows.
Dynamic Dashboards: Build interactive Excel dashboards where your AI updates data, refreshes charts, and modifies layouts based on changing business requirements.
Client Deliverables: Generate professional spreadsheets for clients with branded formatting, complex calculations, and visual elements—without manual Excel work.
Local Development (stdio transport):
{
"mcpServers": {
"excel-stdio": {
"command": "uvx",
"args": ["excel-mcp-server", "stdio"]
}
}
}
Server Deployment (SSE transport):
EXCEL_FILES_PATH=/data/spreadsheets uvx excel-mcp-server sse
Then connect via:
{
"mcpServers": {
"excel": {
"url": "http://localhost:8000/sse"
}
}
}
Built on battle-tested Python libraries—openpyxl for Excel manipulation, pandas for data processing, matplotlib for charts. This isn't a wrapper around Excel; it's a complete reimplementation that's often more reliable than the original.
With 1,200+ stars and active development, this server has proven itself in production environments where traditional Excel automation failed.
Ready to eliminate Excel licensing costs while giving your AI agents spreadsheet superpowers? Install with uvx excel-mcp-server stdio and start building.