Univer MCP

🚀 AI-powered spreadsheet automation through Model Context Protocol (MCP)

Overview

Univer MCP is a Model Context Protocol (MCP) server that enables AI applications to interact with and automate spreadsheet operations using the powerful Univer framework. It bridges the gap between AI language models and spreadsheet functionality, allowing for intelligent spreadsheets processing and automation.

⚠️ Early Stage: Univer MCP is currently in early development. We welcome feedback, suggestions, and collaboration from the community to help shape its future direction.

Key Features

📊 Spreadsheet Operations: Full support for creating, editing, and manipulating spreadsheets
🔧 Extensible Architecture: Built on Univer's plugin system for custom functionality
⚡ High Performance: Leverages Univer's optimized rendering and calculation engines
📦 Advanced Features: Formulas, conditional formatting, data validation, and more
🤖 AI Integration: Seamless integration with MCP-compatible AI applications
📈 Chart Support[WIP]: Creating and updating charts in spreadsheets
🔗 Pivot Table[WIP]: Creating and updating pivot tables in spreadsheets
🌐 Cross-Platform[WIP]: Works across web browsers and Node.js environments
🔄 Real-time Collaboration[WIP]: Support for collaborative editing and real-time updates

https://github.com/user-attachments/assets/c77c7927-335e-47fd-9cf7-852d4880d4b9

Requirements

Multi-modal Model required: Some Univer mcp tools support returning images for better understanding, so the model you choose should ideally support multimodality.

Plain text mode is currently experimental(NOT supported yet); it may be supported in the future. Any suggestions in this regard are very welcome.

How It Works

MCP host: such as cursor, claude code, or the agents you built.
Univer MCP Server: provides the endpoint for univer mcp, which proxies tool calls to the univer instance for execution.
mcp-bridge: this is a univer plugin used to handle various spreadsheet commands in mcp.
univer instance: the runtime environment of univer (spreadsheet).

Quick Start

Get API Key

First, you need to get an API key from the API Keys page. This key will be used to authenticate your MCP server connection.

Start Univer Instance

Before you talk to LLM, you need to launch a univer instance, which is a spreadsheet runtime where you will see the contents of the spreadsheet and how it is being operated.

There are two ways to start a univer instance quickly:

Use the Univer MCP Playground to start a univer instance.
Use the Univer MCP Start-kit to start a univer instance.

Configure MCP Server

Add the following configuration to your MCP client settings:

{
  "mcpServers": {
    "univer": {
      "url": "https://mcp.univer.ai/mcp/?univer_session_id=default",
      "type": "http",
      "headers": {
        "Authorization": "Bearer <API_KEY>"
      }
    }
  }
}

[!TIP] univer_session_id is the session id of your Univer MCP session. Optional, default by default. Keep in mind that the session id MUST be consistent with the session id of your Univer instance.

Cursor

Click the button to install:

Claude Code

claude mcp add --transport http univer-mcp https://mcp.univer.ai/mcp/ -H 'Authorization: Bearer {Your UNIVER_API_KEY}'

Gemini CLI

gemini mcp add --transport http univer-mcp https://mcp.univer.ai/mcp/ --header "Authorization: Bearer {Your UNIVER_API_KEY}"

References

Univer MCP Console - Get your API keys and manage your account
Documentation Guide - Comprehensive usage guide
Playground - Playground for spreadsheet operations
Start Kit - Ready-to-use templates and examples

Resources & Support

Main Project: Univer Framework
Community: Join our Discord for discussions and support
Issues: Report bugs and request features on GitHub Issues
Stack Overflow: Tag your questions with univer-mcp

License

This project is licensed under the MIT License - see the LICENSE file for details.

Made with ❤️ by the DreamNum team