Manage Airtable records with AI
Use the official command-line tool to look up and edit records. Inspect a table’s fields and test the change on one item before applying it to a batch.
Manage Airtable records with AI by making the first request small enough to inspect. A shop might want to change a packaging code for one size of a product. An exact product name can still match three sizes. Reading the matching rows before changing them gives the operator a chance to choose the correct one.
The airtable-cli guide starts with the commands the live service makes available. That matters because access and tool definitions can differ between accounts. The agent can inspect a table, read a field’s actual options and perform a permitted update once the intended record is clear.
What it does
- List the bases available to the authorised account
- Read table fields before constructing a filter
- Find a record by a value in a text field
- Update an agreed item and read its value back
- Use a page-scoped command when access is limited to an interface
Install and use
- Install Airtable’s official MCP CLI from npm in a terminal-enabled agent environment. Configure a personal access token with the required permissions, keeping its value outside the conversation.
- Ask, “Find the large navy storage box. Show its current packaging code and the matching record before making a change.”
- Read back the first update. Check the current batch limit and matching fields before applying the same change to any further items.
FAQ
Source
Made by Airtable. Original skill: airtable-cli, published on GitHub under the MIT licence.
View on GitHub