Daily LinkedIn competitor ad swipe file in Airtable

Every morning at 7am, build and maintain a competitor ad swipe file in Airtable. This should be a deterministic code workflow with no reasoning involved, just a closed pipeline: list, diff, insert.

Trigger: cron, daily at 7am in the user's timezone.

Inputs the user configures once: (1) a list of competitor advertiser URNs to track on LinkedIn, (2) the Airtable base ID and table ID for the swipe file, (3) the name of the field in that table that stores the LinkedIn Ad URN (used as the dedupe key), and (4) the LinkedIn country list to scope the search to (default to US).

For each advertiser URN in the configured list, call LinkedIn Ad Library's Search Ad Library operation, filtered to ads first seen in the last 24 hours and to the configured countries. Paginate through nextPageToken until exhausted.

Before inserting any ad, call Airtable's List Records on the swipe file table with a filterByFormula that matches the ad's URN against the configured Ad URN field. If a record already exists, skip it. This is the dedupe step and it must happen for every ad before any Create Records call.

For each genuinely new ad, call Airtable's Create Records with this fixed schema: Advertiser, Ad URN, Headline, Body, CTA, Landing URL, Country, Date First Seen, Ad Format, Impressions Range. Batch up to 10 inserts per Create Records call to respect Airtable's rate limit of 5 requests per second per base.

Handle the obvious failure modes: skip advertisers that return 0 ads, log and continue on any single-ad failure rather than aborting the run, and back off on 429 responses from either API. The run should be idempotent. If it crashes halfway through and reruns later the same day, it must not insert duplicates.

The point of this workflow is that a designer or strategist opens Airtable on Friday with a full week of new competitor creative already organized and filterable. No notifications, no AI commentary, just a clean swipe file that maintains itself.