Excel

You can run the Excel MCP Server locally to analyze Excel and CSV data with a built‑in Excel-like formula engine, AI-assisted insights, and multi-file processing. It exposes practical tools for reading, querying, cleaning, and aggregating data across files, all accessible from natural language conversations.

How to use

Activate the MCP server in your environment and connect it to your MCP client (Claude or equivalent). You can run the server locally and then issue natural language queries like asking for totals, averages, or cross-file checks. Use standard file reading commands and the built‑in tools to read CSV/Excel files, evaluate formulas, and perform bulk operations across multiple files. When you ask questions in natural language, the system will generate and execute appropriate formulas, perform data analysis, and return results in a familiar spreadsheet-like structure.

How to install

Prerequisites you need before installation: Node.js and npm installed on your machine.

Install and build the MCP server locally using the following steps.

git clone https://github.com/ishayoyo/excel-mcp.git
cd excel-mcp
npm install
npm run build

Configuration and usage notes

You can start using the server through either of these local, self‑contained MCP configurations.

// Option 1: Start via Node pointing to the built entry
command: node
args: ["C:/path/to/excel-mcp/dist/index.js"]

// Option 2: Start via npx (global install)
command: npx
args: ["-y", "excel-csv-mcp"]

AI providers and environment setup

Set up your AI provider keys in a standard environment file so the server can route natural language requests to the best available AI service.

ANTHROPIC_API_KEY=your-key-here
OPENAI_API_KEY=your-key-here
DEEPSEEK_API_KEY=your-key-here
GEMINI_API_KEY=your-key-here

Available tools

read_file

Read CSV/Excel files with offset/limit support for large datasets.

read_file_chunked

Stream large files in chunks to handle datasets bigger than memory limits.

get_file_info

Analyze a file’s size and suggest optimal chunking and access strategies for large datasets.

get_cell

Access a single cell value by row/column or address.

get_range

Retrieve a range of cells or a block of data.

search

Search for data matching criteria across files.

filter_rows

Filter rows based on given conditions.

aggregate

Perform aggregations like SUM, AVG, MIN, MAX, COUNT.

find_duplicates

Identify duplicate rows or values across files.

data_cleaner

Batch data cleaning tasks such as trimming spaces and standardizing formats.

vlookup_helper

Assist with VLOOKUP-style lookups including fuzzy matching and error handling.

bulk_aggregate_multi_files

Aggregate data across multiple files in parallel for faster results.

bulk_filter_multi_files

Filter data across multiple files simultaneously.

evaluate_formula

Evaluate Excel-like formulas with dependency tracking.

parse_natural_language

Convert natural language questions into executable formulas.

explain_formula

Explain what a given formula does in plain language.

statistical_analysis

Perform comprehensive statistics on numeric data.

correlation_analysis

Find relationships between numeric columns.

data_profile

Generate a complete data quality and profiling report.

pivot_table

Create pivot-style aggregations and groupings.

validate_data_consistency

Check cross-file data integrity and referential consistency.

write_file

Create and export files with results and formatting.

export_analysis

Export analysis results to preferred formats.