mcp-server-excel
Control Excel with Natural Language through AI assistants like GitHub Copilot, Claude, and ChatGPT. This MCP server enables AI-powered Excel automation for Power Query, DAX measures, VBA macros, PivotTables, and more.
README
ExcelMcp - MCP Server for Microsoft Excel
Automate Excel with AI - A Model Context Protocol (MCP) server for comprehensive Excel automation through conversational AI.
Talk to Excel using GitHub Copilot or Claude to automate Power Query M code, Power Pivot DAX measures, PivotTables, Excel Tables, formatting, data transformations, and complex analysis - all through natural language commands. No programming required. Built on Excel's native COM API for zero risk of file corruption.
🤔 What is This?
ExcelMcp connects AI assistants (GitHub Copilot, Claude) to Microsoft Excel for hands-free automation of complex Excel tasks.
Stop manually clicking through Excel menus for repetitive tasks. Instead, describe what you want in plain English:
Data Transformation & Analysis:
- "Optimize all my Power Queries in this workbook for better performance"
- "Create a PivotTable from SalesData table showing top 10 products by region with sum and average"
- "Build a DAX measure calculating year-over-year growth with proper time intelligence"
Formatting & Styling (No Programming Required):
- "Format the revenue columns as currency, make headers bold with blue background, and add borders to the table"
- "Apply conditional formatting to highlight values above $10,000 in red and below $5,000 in yellow"
- "Convert this data range to an Excel Table with style TableStyleMedium2, add auto-filters, and create a totals row"
Workflow Automation:
- "Find all cells containing 'Q1 2024' and replace with 'Q1 2025', then sort the table by Date descending"
- "Add data validation dropdowns to the Status column with options: Active, Pending, Completed"
- "Merge the header cells, center-align them, and auto-fit all column widths to content"
The AI assistant analyzes your request, generates the proper Excel automation commands, and executes them directly in your Excel application - no formulas or programming knowledge required.
Real-World Example - Power Query Optimization:
You: "This Power Query is taking 5 minutes to refresh. Can you optimize it?"
AI Assistant (using ExcelMcp):
1. Exports the M code to analyze performance bottlenecks
2. Identifies queries that can use query folding
3. Refactors the M code to push operations to the data source
4. Updates the query in your workbook
5. Tests the refresh (now completes in 30 seconds)
Result: A professionally optimized Power Query with documented improvements
🛡️ 100% Safe - Uses Excel's Native COM API
Unlike third-party libraries that manipulate .xlsx files directly (risking file corruption), ExcelMcp uses Excel's official COM API. This ensures:
- ✅ Zero risk of document corruption - Excel handles all file operations safely
- ✅ Interactive development - See changes in real-time as you work with live Excel files
- ✅ Timeout protection - Built-in safeguards prevent indefinite hangs (2-min default, 5-min max)
- ✅ Comprehensive automation - Currently supports 166 operations across 11 specialized tools covering Power Query, Data Model/DAX, VBA, PivotTables, Excel Tables, ranges, and more
🔧 Advanced Features:
- Batch Operations - Group multiple operations in a single Excel session for 75-90% faster execution
- Timeout Protection - Automatic timeout enforcement with rich error messages and LLM-friendly guidance (docs)
- Progress Logging - Real-time operation status updates via stderr (MCP protocol compatible)
- Error Recovery - Intelligent retry suggestions and operation-specific troubleshooting guidance
👥 Who Should Use This?
Perfect for:
- ✅ Data analysts automating repetitive Excel workflows
- ✅ Developers building Excel-based data solutions
- ✅ Business users managing complex Excel workbooks
- ✅ Teams maintaining Power Query/VBA/DAX code in Git
Not suitable for:
- ❌ Server-side data processing (use libraries like ClosedXML, EPPlus instead)
- ❌ Linux/macOS users (Windows + Excel installation required)
- ❌ High-volume batch operations (consider Excel-free alternatives)
🎯 What You Can Do
Development & Automation:
- 🔄 Power Query - 16 operations: atomic workflows (create, update-and-refresh, refresh-all), manage transformations, load configurations (worksheet, data model, connection only), error diagnostics, M code evaluation
- 📊 Power Pivot (Data Model) - 15 operations: build DAX measures, manage relationships, discover model structure (tables, columns), export to .dax files
- 🎨 Excel Tables - 26 operations: automate formatting, filtering, sorting, structured references, number formats, column management
- 📈 PivotTables - 20 operations: create and configure PivotTables for interactive analysis
- 📝 VBA Macros - 7 operations: export/import/run VBA code, integrate with version control
- 📋 Ranges & Data - 45 operations: values, formulas, copy/paste, find/replace, formatting, validation, merge, conditional formatting, cell protection
- 📄 Worksheets - 13 operations: lifecycle management, tab colors, visibility controls
- 🔌 Connections - 11 operations: manage OLEDB, ODBC, Text, Web data sources
- 🏷️ Named Ranges - 7 operations: named range management and bulk operations
<details> <summary>📚 <strong>See Complete Feature List (100+ Operations)</strong></summary>
Power Query & M Code (16 operations)
✨ NEW: Atomic Operations - Single-call workflows replace multi-step patterns:
- Create - Import + load in one operation (replaces import → load workflow)
- Update & Refresh - Update M code + refresh data atomically
- Refresh All - Batch refresh all queries in workbook
- Update M Code - Stage code changes without refreshing data
- Unload - Convert loaded query to connection-only
Core Operations:
- Create, read, update, delete Power Query transformations
- Export/import M code for version control
- Manage query load destinations (worksheet/data model/connection-only/both)
- Set privacy levels for data source combinations
- Get load configuration for existing queries
- Error diagnostics and M code evaluation
- List Excel workbook sources for Power Query integration
Data Model & DAX (Power Pivot) (15 operations)
- Create/update/delete DAX measures with format types (Currency, Percentage, Decimal, General)
- Manage table relationships (create, toggle active/inactive, delete)
- Discover model structure (tables, columns, measures, relationships)
- Export measures to .dax files for Git workflows
- Get comprehensive model information
- Note: DAX calculated columns are not supported (use Excel UI for calculated columns)
Excel Tables (ListObjects) (26 operations)
- Lifecycle: create, resize, rename, delete, get info
- Styling: apply table styles, toggle totals row, set column totals
- Column management: add, remove, rename columns
- Data operations: append rows, apply filters (criteria/values), clear filters, get filter state
- Sorting: single-column sort, multi-column sort (up to 3 levels)
- Number formatting: get/set column number formats
- Advanced features: structured references, Data Model integration
PivotTables (20 operations)
- Creation: create from ranges or Excel Tables
- Field management: add/remove fields to Row, Column, Value, Filter areas
- Aggregation functions: Sum, Average, Count, Min, Max, etc. with validation
- Advanced features: field filters, sorting, custom field names, number formatting
- Data extraction: get PivotTable data as 2D arrays for further analysis
- Lifecycle: list, get info, delete, refresh
VBA Macros (7 operations)
- List all VBA modules and procedures
- View module code without exporting
- Export/import VBA modules to/from files
- Update existing modules
- Execute macros with parameters
- Delete modules
- Version control VBA code through file exports
Ranges & Worksheets
- Data Operations (10 actions): get/set values/formulas, clear (all/contents/formats), copy/paste (all/values/formulas), insert/delete rows/columns/cells, find/replace, sort
- Number Formatting (3 actions): get formats as 2D arrays, apply uniform format, set individual cell formats
- Visual Formatting (1 action): font (name, size, bold, italic, underline, color), fill color, borders (style, weight, color), alignment (horizontal, vertical, wrap text, orientation)
- Data Validation (3 actions): add validation rules (dropdowns, number/date/text rules), get validation info, remove validation
- Hyperlinks (4 actions): add, remove, list all, get specific hyperlink
- Smart Range Operations (3 actions): UsedRange, CurrentRegion, get range info (address, dimensions, format)
- Merge Operations (3 actions): merge cells, unmerge cells, get merge info
- Auto-Sizing (2 actions): auto-fit columns, auto-fit rows
- Conditional Formatting (2 actions): add conditional formatting, clear conditional formatting
- Cell Protection (2 actions): set cell lock status, get cell lock status
- Formatting & Styling (3 actions): get style, set style, format range
- 45 range operations total covering all common Excel range manipulation needs
- Worksheet management (13 actions): lifecycle (create, rename, copy, delete), tab colors (set, get, clear), visibility controls (show, hide, very-hide, get/set status)
Data Connections (11 operations)
- Manage OLEDB, ODBC, Text, Web connections
- Import/export connections via .odc files
- Update connection strings and properties
- Refresh connections and test connectivity
- Load connection-only connections to worksheet tables
- Get/set connection properties (refresh settings, background query, etc.)
Named Ranges (7 operations)
- List all named ranges with references
- Get/set single values
- Create/delete named ranges
- Update cell references
- Bulk create multiple parameters
</details>
🚀 Quick Start (2 Minutes)
Requirements: Windows OS + Microsoft Excel 2016+
⚠️ Important: Close all Excel files before using ExcelMcp. The server requires exclusive access to workbooks during automation (Excel COM limitation).
⭐ Recommended: VS Code Extension (One-Click Setup)
Fastest way to get started - everything configured automatically:
-
Install Extension
- Open VS Code → Extensions (
Ctrl+Shift+X) - Search for "ExcelMcp"
- Click Install
- Open VS Code → Extensions (
-
Automatic Setup (no manual steps!)
- ✅ Installs .NET 8 runtime
- ✅ Includes bundled MCP server
- ✅ Registers with AI assistants
- ✅ Shows quick start guide
-
Start Using It
The extension opens automatically after installation with a quick start guide!
Manual Installation (Advanced Users)
For non-VS Code environments or manual setup:
# Install .NET 8 SDK
winget install Microsoft.DotNet.SDK.8
# Install ExcelMcp MCP server as a global tool
dotnet tool install --global Sbroenne.ExcelMcp.McpServer
# To update to the latest version later:
dotnet tool update --global Sbroenne.ExcelMcp.McpServer
Configure Your AI Assistant
For GitHub Copilot in VS Code - Create .vscode/mcp.json in your workspace:
{
"servers": {
"excel": {
"command": "dotnet",
"args": ["tool", "run", "mcp-excel"]
}
}
}
For GitHub Copilot in Visual Studio - Create .mcp.json in your solution directory or %USERPROFILE%\.mcp.json:
{
"servers": {
"excel": {
"command": "dotnet",
"args": ["tool", "run", "mcp-excel"]
}
}
}
For Claude Desktop - Add to your MCP configuration:
{
"mcpServers": {
"excel": {
"command": "dotnet",
"args": ["tool", "run", "mcp-excel"]
}
}
}
Test It Out
Try a practical example - ask your AI assistant:
Create an empty Excel file called "test.xlsx" and add a Power Query that loads data from a CSV file
The AI will guide you through the process and execute the commands directly!
🔧 How It Works - COM Interop Architecture
ExcelMcp uses Windows COM automation to control the actual Excel application (not just .xlsx files).
This means you get:
- ✅ Full Excel Feature Access - Power Query engine, VBA runtime, Data Model, calculation engine, pivot tables
- ✅ True Compatibility - Works exactly like Excel UI, no feature limitations
- ✅ Live Data Operations - Refresh Power Query, connections, Data Model in real workbooks
- ✅ Interactive Development - Immediate Excel feedback as AI makes changes
- ✅ All File Formats - .xlsx, .xlsm, .xlsb, even legacy formats
Technical Requirements:
- ⚠️ Windows Only - COM interop is Windows-specific
- ⚠️ Excel Required - Microsoft Excel 2016 or later must be installed
- ⚠️ Desktop Environment - Controls actual Excel process (not for server-side processing)
🔟 MCP Tools Overview
11 specialized tools for comprehensive Excel automation:
- excel_powerquery (16 actions) - Power Query M code: create, view, import, export, update, delete, manage load destinations, privacy levels, errors, eval, list Excel sources
- excel_datamodel (15 actions) - Power Pivot (Data Model): CRUD DAX measures/relationships, discover structure (tables, columns), export to .dax files
- excel_table (26 actions) - Excel Tables: lifecycle, columns, filters, sorts, structured references, totals, number formatting, Data Model integration
- excel_pivottable (20 actions) - PivotTables: create from ranges/tables, field management (row/column/value/filter), aggregations, filters, sorting, extract data
- excel_range (45 actions) - Ranges: get/set values/formulas, number formatting, visual formatting (font, fill, border, alignment), data validation, clear, copy, insert/delete, find/replace, sort, hyperlinks, merge, conditional formatting, cell protection
- excel_vba (7 actions) - VBA: list, view, export, import, update, run, delete modules
- excel_connection (11 actions) - Connections: OLEDB/ODBC/Text/Web management, properties, refresh, test
- excel_worksheet (13 actions) - Worksheets: lifecycle (list, create, rename, copy, delete), tab colors (set-tab-color, get-tab-color, clear-tab-color), visibility (set-visibility, get-visibility, show, hide, very-hide)
- excel_namedrange (7 actions) - Named ranges: list, get, set, create, create-bulk, delete, update
- excel_file (3 actions) - File operations: create empty .xlsx/.xlsm workbooks, close workbook, test
- excel_batch (3 actions) - Multi-operation performance: begin, commit, list
Total: 11 tools with 166 actions
📚 Complete MCP Server Guide → - Detailed tool documentation and examples
📋 Additional Information
Testing Philosophy
Why No Unit Tests? ExcelMcp uses integration tests exclusively because Excel COM cannot be meaningfully mocked. Our integration tests ARE our unit tests. See ADR-001: No Unit Tests for full architectural rationale.
CLI for Direct Automation
ExcelMcp also provides a command-line interface for vba-based Excel automation (no AI required). See CLI Guide for complete documentation.
Project Information
License: MIT License - see LICENSE file
Contributing: See CONTRIBUTING.md for guidelines
Built With: This entire project was developed using GitHub Copilot AI assistance
Acknowledgments:
- Microsoft Excel Team - For comprehensive COM automation APIs
- Model Context Protocol community - For the AI integration standard
- Open Source Community - For inspiration and best practices
SEO & Discovery
Excel Automation • Automate Excel with AI • MCP Server • Model Context Protocol • GitHub Copilot Excel • AI Excel Assistant • Power Query Automation • Power Query M Code • Power Pivot Automation • DAX Measures • DAX Automation • Data Model Automation • PivotTable Automation • VBA Automation • Excel Tables Automation • Excel AI Integration • COM Interop • Windows Excel Automation • Excel Development Tools • Excel Productivity • Excel Scripting • Conversational Excel • Natural Language Excel
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。