← Back to directory
E

Excel MCP Server

Community
An MCP server for manipulating Excel files without Microsoft Excel
Category
Filesystem #22 of 24
Stars
★ 4.2k Very popular
Transport
stdio (local process) · SSE · Streamable HTTP
Runtime
Python
Credentials
No credential needed
License
MIT
Last commit
41FMRS · D

The project provides an MCP service for Excel file manipulation with local and remote deployment options and three transports. The cached source does not include a server manifest or TOOLS.md, so specific tool identifiers and client compatibility levels cannot be verified.

Strongest · Setup experience 12/20 Weakest · Reliability 5/20

Reliability
5/20
Security and permissions
7/20
Maintenance
7/20
Documentation
10/20
Setup experience
12/20
Why each score
Reliability 5/20
The README describes stdio, SSE, and streamable HTTP entry points and basic connection examples, making the happy path plausible. However, there is no server manifest, source code, CI, test suite, or implementation evidence to verify MCP initialization, tool listing, dependencies, error handling, or whether declared behavior matches reality, so the score is conservative.
Security and permissions 7/20
The README states that HTTP transports reject absolute paths and directory traversal and use EXCEL_FILES_PATH as a file-root boundary, which is explicit scoping evidence. However, stdio accepts client-supplied paths, while authentication, authorization, TLS, tenant isolation, confirmation for workbook writes, and the data-flow model for remote use are not documented. No red-line risk is evidenced by the supplied material, but these gaps warrant deductions.
Maintenance 7/20
The repository is not archived, has an MIT license, and shows substantial public interest and open issue activity; these are not sufficient evidence of sustained maintenance. No release history, commit activity, dependency update process, issue-response evidence, maintainer identity, or security-response channel is supplied, so only a low-to-moderate score is justified.
Documentation 10/20
The README includes installation commands, three transport configurations, client examples, port and file-path environment variables, and a feature overview with a pointer to TOOLS.md. It does not provide the actual tool parameter reference, authentication guidance, secure deployment instructions, limits, cost information, compatibility details, troubleshooting, or verifiable implementation evidence; the manifest is also unavailable.
Setup experience 12/20
The README provides uvx launch commands and JSON client configurations for local stdio and remote HTTP, making the basic setup relatively clear. It assumes uv/uvx, omits remote deployment, authentication, TLS, and platform-specific guidance, and supplies no verifiable CI or tests for the connection path, so it does not merit a near-perfect score.

Static review · not runListed 2026-07-29

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesWrites / deletes local files

Best for

  • AI agent workflows that automate Excel file handling
  • Local or remote services that should not depend on Microsoft Excel installation
  • Users who need Excel data and formatting operations through MCP

Not for

  • Workflows requiring the Microsoft Excel desktop application itself
  • Use cases requiring a verified complete tool inventory when TOOLS.md is not provided
  • Deployments where SSE is intended as the preferred transport

Required permissions

  • In stdio mode, the client supplies the Excel file path for each tool call
  • In SSE and streamable HTTP modes, the server must define the Excel file root through `EXCEL_FILES_PATH`
  • In SSE and streamable HTTP modes, file paths must be relative to that root directory

Risks and side effects

  • The server can create, read, and modify workbooks, so incorrect operations may change file contents
  • SSE is marked as deprecated in the project documentation
  • Remote file access is constrained by the `EXCEL_FILES_PATH` root and relative-path rules

Setup

Before you start

Runtime:Python

Other optional settings (2)
EXCEL_FILES_PATH optional Directory the server reads/writes Excel files in for SSE/streamable-http transports; defaults to ./excel_files, not needed for stdio.
FASTMCP_PORT optional Port the server listens on in SSE/streamable-http transports; defaults to 8017.

For local stdio use, run uvx excel-mcp-server stdio and use the provided MCP client configuration. For SSE, run uvx excel-mcp-server sse and connect to http://localhost:8000/sse; this transport is deprecated. For remote connections, the README recommends uvx excel-mcp-server streamable-http with the endpoint http://localhost:8000/mcp. SSE and streamable HTTP require the server-side EXCEL_FILES_PATH environment variable. Set FASTMCP_PORT to change the listening port; the default is 8017.

MCP config · JSON
{"mcpServers":{"excel":{"command":"uvx","args":["excel-mcp-server","stdio"]}}}

Works in clients that read an "mcpServers" key. For clients such as VS Code that use a different key, the configurator below converts it.

.vscode/mcp.json
{
  "servers": {
    "excel": {
      "command": "uvx",
      "args": [
        "excel-mcp-server",
        "stdio"
      ]
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal
claude mcp add excel -- uvx excel-mcp-server stdio

Run it in a terminal; replace any <…> placeholders with your own values first.

Check that it works

After adding the excel server to your client, Excel manipulation tools should appear in the tool list; you can also ask it to create a test workbook to confirm the connection.

Troubleshooting

  1. Verify that the server-side `EXCEL_FILES_PATH` is set for SSE or streamable HTTP
  2. Ensure remote `filepath` values are relative and do not use absolute paths or directory traversal
  3. Verify that the endpoint matches the transport: `/sse` for SSE and `/mcp` for streamable HTTP
  4. Check `FASTMCP_PORT` when using a non-default port; the default is 8017
  5. For stdio, verify that each tool call supplies the correct file path

Things to try

Once connected, you can ask your AI assistant things like:

  • Create a new Excel workbook and add a sheet called Sales Data
  • Read the data from the first sheet of reports/q1.xlsx
  • Add a line chart of monthly revenue to sales.xlsx
  • Copy Sheet1, rename the copy to Backup, then delete the original sheet

Use cases

Creating, reading, and updating Excel workbooks and worksheets
Applying formulas, formatting, charts, pivot tables, and Excel tables
Using an AI agent to automate Excel data operations

Overview

excel-mcp-server is a Model Context Protocol server for creating, reading, and modifying Excel workbooks and worksheets without Microsoft Excel installed. It supports formulas, formatting, charts, pivot tables, Excel tables, data validation, and worksheet management over stdio, SSE, or streamable HTTP.

Similar servers

Source revision 711f07c99240 Data synced 2026-10-19 Read the FMRS scoring method