Airtable agent

The Airtable agent is a specialized agent node that enables your workflows to interact directly with Airtable databases via APIs. It allows querying, updating, appending, and reading records from your Airtable bases in a conversational or programmatic manner, leveraging both automation and AI.

What is the Airtable agent?

• Type: Standalone agent node within the platform, specialized for Airtable workflows.
• Category: External Integration, Database agent
• Icon: (if image supported in GitBook build)

Use Cases

• Integrate Airtable data into your AI assistant ("show me all open support tickets").
• Implement automations that update, append, or modify Airtable rows from chatflows.
• Automate report generation, cross-database lookups, and more using AI or workflow triggers.
• Build Slackbots, chatbots, or forms that read/write to Airtable.

Inputs/Outputs & Key Configuration

Inputs

• API Key / OAuth Credentials: Required to access your Airtable workspace. Store these in Credentials and reference in the node.
• Table & Base Selection: Specify the Base ID, Table Name or ID you wish to operate on.
• Action: Choose from supported operations (query, append, update, read).
• Query/Filter: For read/query—specify Airtable filter formula, sort, or view.
• Record Data: For append/update—provide object/fields for the operation.
• Variables: Can be dynamically injected from prior nodes or user input for fully automated flows.

Outputs

• Data: Returns structured JSON, typically an array of matching Airtable records, or status confirmation on update/append.
• Status/Error: Returns appropriate error message for invalid credentials, query, or write failures.

Example Workflow

Read Support Tickets From Airtable

1. Start Node: Trigger on user message or scheduled time.
2. Airtable agent Node:
   • Select your credential.
   • Set to query on SupportTickets table.
   • Query: Status = 'Open'
3. LLM Node: Summarizes or presents the results.
4. Chat Output: Returns results to user in tabular or summarized form.

Update a Row

1. Start Node: Trigger from webhook/user action.
2. Airtable agent Node:
   • Action: update
   • Record ID: pass as variable from trigger
   • Fields: { Status: 'Closed', ClosedAt: {{now}} }
3. Output: Returns status, which can be used for notifications.

Configuration in UI

• Drag "Airtable agent" node to canvas.
• Open its config:
  1. Select credential (API key / OAuth).
  2. Provide Base ID and Table Name.
  3. Choose Action: Query/Append/Update/Read.
  4. Enter relevant parameters, e.g., filter, record data.
  5. Optionally connect to previous node outputs or variables.
• Connect to downstream nodes (e.g., LLM, Output, Webhook).

Best Practices

• Do not hardcode sensitive keys—use platform credential manager.
• For complex queries, build and test your Airtable formula in Airtable UI first.
• Handle error outputs using condition nodes for robust automations.
• Use variable injection for dynamic table selection, i.e., support multi-tenant/multi-table automations.

Troubleshooting

• 403 errors: check that credentials and permissions are correct.
• Empty results: verify Base ID, Table Name, and query filter.
• Write failures: confirm required fields and data types in Airtable.

For detailed API limits, field types, and formulas, refer to Airtable API Docs.