I built an Airtable connector for Cognee. The hard part was knowing what to delete.

Description and use

An Airtable connector for Cognee was built, implementing full source reconciliation with incremental document emission and deletion markers, correctly handling comment updates and schema changes.

Problem

When syncing Airtable records to Cognee, it is unclear whether to delete obsolete data or keep data that failed to sync, leading to either data loss or stale information.

Outcome

The connector successfully reconciles Airtable data with Cognee, correctly handling deletions, failed syncs, and comment updates, as verified by 107 passing local tests.

Workflow

• 1. Add Airtable connector to cognee-community without changing core.
• 2. Configure connector with primary key, merge write disposition, and unlimited read.
• 3. Perform full source reconciliation: read all selected tables, compare with persisted fingerprints.
• 4. Emit changed documents and confirmed deletion markers.
• 5. Handle comments and schema independently of timestamps.
• 6. On sync failure, retain last successful data; on record deletion, remove corresponding documents and schema.
• 7. Run local verification tests to ensure assertions pass.

Evidence

The connector turns selected records into documents, including comments when enabled.

The design is full source reconciliation with incremental document emission.

107 passing tests on Python 3.12, with no failures, errors, or skips.