domdomegg/airtable-mcp-server

Overview

Connects your AI assistant to Airtable so it can read, create, update, and delete records in your bases, useful for any business that runs projects, inventory, or CRM-style tracking through Airtable. Ask it to look up a record, log a new entry, or update a status, all in plain English. Widely used and actively maintained, making it a dependable choice for teams already on Airtable.

Best for

Best for
Teams running projects, inventory, or CRM tracking in Airtable who want faster updates

Use cases

• Look up a record in an Airtable base from a plain-English request
• Log a new entry or update a status without opening Airtable
• Query a table to answer a question about current data
• Automate routine data management tasks in Airtable

How to use

Install

MCP_TRANSPORT=http PORT=3000 npx airtable-mcp-server

Tools exposed

• list_records
• search_records
• list_bases
• list_tables
• describe_table
• get_record
• create_record
• update_records
• delete_records
• create_table
• update_table
• create_field
• update_field
• create_comment
• list_comments

Tested with

Claude Code, Claude Desktop, Cursor, Cline, VS Code

Example client config

{"AIRTABLE_API_KEY":"pat123.abc123"}

Notes

An open-source Model Context Protocol server that translates MCP requests into Airtable API calls. It allows AI assistants to read, create, update, and delete records in Airtable bases using a standardized protocol. Written in TypeScript and hosted on GitHub with over 440 stars.

443 stars on GitHub. Last updated 2026-05-23. Licensed MIT.

Use cases

• Query Airtable tables from an AI chatbot
• Create and update Airtable records via natural language commands
• Automate data management tasks by connecting AI agents to Airtable

Pros

• Open source and written in TypeScript, easy to audit and extend
• Leverages the Model Context Protocol for standardized AI tool integration
• Simple setup for developers already using Airtable and MCP clients

Cons

• Requires a running MCP client and an Airtable API key
• Limited by Airtable’s API rate limits and base structure
• Not a standalone service; needs user-managed infrastructure or hosting

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.