santoshray02/csv-editor

Stateful CSV editing MCP server for AI assistants — sessions, undo/redo, auto-save, and 39 pandas-powered tools. Works with Claude, ChatGPT, Cursor, Windsurf, Claude Code, and any MCP-compatible client. Built on FastMCP 3

24 ★5 forksPythonUpdated 2mo ago

What you need to know

A stateful CSV editing MCP server with 39 tools for load, transform, clean, analyze, validate and export CSV data, including sessions, undo/redo and auto-save.

Install

npx -y @smithery/cli install @santoshray02/csv-editor --client claude
uv tool install csv-editor
pip install csv-editor

Usage

•Load, filter, transform and export CSV data with natural language
•Add sessions with undo/redo and auto-save

Key features

✓39 tools over CSV data (load, transform, analyze, validate)
✓Stateful sessions, multi-session isolation, undo/redo
✓Auto-save strategies (overwrite/backup/versioned/custom)
✓I/O for CSV, JSON, Excel, Parquet, HTML, Markdown
✓Quality scoring, schema validation, anomaly detection

Best for

AI agents that need to edit and iterate on CSV files, not just read them

Caveats

⚠Best for files under 1 GB (pandas-based)
⚠Remote HTTP/OAuth for ChatGPT Connectors is on the roadmap, not yet shipped

Platforms: macOS · Windows · LinuxClients: Claude Desktop · Claude Code · Cursor · Windsurf · VS Code Copilot · Cline · Continue · Zed

Reviewed 2026-08-11

Topics

csv data-analysisdata-cleaningdata-manipulationdata-qualitydata-validationetl fastmcpllmmcpmcp-servermodel-context-protocolpandaspython3

Stars: 24★
Forks: 5
Language: Python
License: MIT
Created: 2025-08-12
Last push: 2026-07-14