n8n Workflow: Facebook Ad Angle Classifier w/ Apify & OpenAI

Turn Facebook Ad Library data into searchable, angle-labeled intel—automatically

This n8n Workflow scrapes ads from the Facebook Ad Library using Apify, normalizes destinations, then uses OpenAI to classify each ad’s primary persuasion angle. It stores enriched results in Supabase and embeds ad copy for semantic search.

What this workflow does

• Manual run setup: You start the workflow and provide a Facebook Ad Library search URL and the maximum number of ads to scrape.
• Scrape ads via Apify: The workflow runs the Apify actor curious_coder/facebook-ads-library-scraper to collect ad details (including ad copy and creatives, plus metadata such as reach/dates).
• Normalize & extract destinations: Each ad is converted into a consistent row, Facebook redirect links are unwrapped to the real destination, and a destination domain is derived. It also builds a Shopify probe URL.
• Classify persuasion angle with OpenAI: For each ad, the workflow sends its text to OpenAI to label its primary persuasion angle (e.g., urgency, social proof, or problem solution).
• Store results in Supabase: Inserts enriched ads into the Supabase ads table. For rows inserted successfully, it embeds the ad copy into a Supabase documents vector store with metadata.
• Optional Shopify product enrichment: If a valid Shopify probe URL exists, the workflow requests /products.json (or /products/<handle>.json), deduplicates products by product ID, and inserts them into a Supabase products table.

Use cases

• Competitive ad research: Track which persuasion angles competitors use across campaigns found in the Facebook Ad Library.
• Creative strategy testing: Build a searchable library of ad copy and angle labels to inform new messaging.
• Shopify-linked product discovery: Enrich ads with Shopify product data to connect creative themes to catalog items.

Technical details

• n8n nodes: set, code, filter, Supabase, sticky note, HTTP Request.
• Integrations: Apify (with curious_coder/facebook-ads-library-scraper), OpenAI (angle classification + embeddings), and Supabase (tables + vector store).