Excel MCP Server
An MCP server for manipulating Excel files that features a headless engine for real-time formula calculation and data validation. It enables users to create, read, and manage workbooks, sheets, and charts without requiring Microsoft Excel installed.
README
Excel MCP Server
A Model Context Protocol (MCP) server for Excel file manipulation with real-time formula calculation.
Why a Formula Engine?
Problem: Both humans and LLMs are inconsistent at math. When an AI writes =SUM(A1:A10) to a spreadsheet, it's just storing a string—the calculation never happens unless Microsoft Excel opens the file.
Solution: This MCP integrates HyperFormula, a headless spreadsheet engine that evaluates formulas in real-time. When you read a cell, you get the calculated result, not just the formula text.
write_cell(wb, "A1", 100)
write_cell(wb, "A2", 200)
write_cell(wb, "A3", "=SUM(A1:A2)")
read_cell(wb, "A3")
→ { calculatedValue: 300, formula: "=SUM(A1:A2)" }
This ensures mathematical accuracy for:
- Financial calculations
- Data aggregations
- Report generation
- Any workflow where precision matters
Features
- No Excel Required - Works with .xlsx files using pure JavaScript
- 400+ Excel Functions - SUM, AVERAGE, VLOOKUP, IF, and more via HyperFormula
- Multi-Workbook - Open and manipulate multiple workbooks simultaneously
Tools (21 total)
Workbook Management
| Tool | Description |
|---|---|
create_workbook |
Create new workbook in memory, returns ID |
open_workbook |
Open existing .xlsx file |
save_workbook |
Save workbook to disk |
close_workbook |
Close workbook and free memory |
list_workbooks |
List all open workbooks |
Sheet Operations
| Tool | Description |
|---|---|
list_sheets |
List all sheets in workbook |
create_sheet |
Add new sheet |
delete_sheet |
Remove sheet |
rename_sheet |
Rename sheet |
get_sheet_info |
Get dimensions and metadata |
Cell Operations
| Tool | Description |
|---|---|
read_cell |
Read cell value (returns calculated result) |
write_cell |
Write value or formula (prefix with =) |
read_range |
Read range of cells |
write_range |
Write 2D array to range |
list_columns |
List columns with headers |
Formula Support
| Tool | Description |
|---|---|
get_formula |
Get raw formula string from cell |
recalculate |
Force recalculation of all formulas |
Charts & Pivot Tables
| Tool | Description |
|---|---|
create_chart |
Create bar, line, pie, or scatter chart |
delete_chart |
Remove chart |
create_pivot_table |
Create pivot table from data |
Import
| Tool | Description |
|---|---|
import_markdown_table |
Import markdown table to Excel |
Installation
git clone https://github.com/jauinones/excel-mcp.git
cd excel-mcp
npm install
npm run build
Configuration
Add to your Claude Code MCP settings (~/.claude.json):
{
"mcpServers": {
"excel": {
"type": "stdio",
"command": "node",
"args": ["/path/to/excel-mcp/dist/index.js"]
}
}
}
Example Usage
// Create a simple invoice
create_workbook → "wb_1"
write_cell(wb_1, "Sheet1", "A1", "Item")
write_cell(wb_1, "Sheet1", "B1", "Price")
write_cell(wb_1, "Sheet1", "C1", "Qty")
write_cell(wb_1, "Sheet1", "D1", "Total")
write_range(wb_1, "Sheet1", "A2", [
["Widget", 25.00, 10, "=B2*C2"],
["Gadget", 15.50, 20, "=B3*C3"],
["Thing", 8.99, 50, "=B4*C4"]
])
write_cell(wb_1, "Sheet1", "D5", "=SUM(D2:D4)")
read_cell(wb_1, "Sheet1", "D5")
// → { calculatedValue: 1009.5, formula: "=SUM(D2:D4)" }
save_workbook(wb_1, "/path/to/invoice.xlsx")
Tech Stack
- TypeScript - Type-safe implementation
- ExcelJS - Read/write .xlsx files
- HyperFormula - Formula calculation engine (GPLv3)
- MCP SDK - Model Context Protocol integration
License
GPLv3 (due to HyperFormula dependency)
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。