Agent recipes

Analyze a competitor#

1. lookup(q="gymshark.com") → advertiser id and shop domain
2. get_advertiser(id) → activity, platforms, formats
3. get_advertiser_ads(id, active=true, sort="longest_running") → the winners
4. get_shop(domain) and get_similar_shops(domain) → stack and competitors

Find winning products in a niche#

1. list_facets(facet="niches") → pick a slug
2. search_products(preset="new_winners", niches=["pet-supplies"])
3. For each product, search_ads(query=<title>, min_running_days=30) to confirm demand
4. get_similar_ads(id) → other sellers of the same product

Build a swipe file#

1. search_ads(query="before after", formats=["video"], active=true, min_running_days=14)
2. create_board(name="Before/after hooks")
3. save_ad_to_board(board_id, ad_id) for the best ones

Weekly competitor digest#

1. list_trackers()
2. get_tracker_ads(id, sort="newest", limit=10) for each
3. Summarize new angles and offers per brand