Competitor Ad Research Pipeline

Multi-tenant competitor ad research system: scrape Meta Ads Library, filter by longevity, AI-analyze with two-pass pipeline, store in PostgreSQL, collect client feedback via interactive rating page, feed insights into content strategy.

Architecture

Apify Scraper (Meta Ads Library)
  → Longevity Filter (14d+ = proven)
  → Pass 1: Haiku 4.5 (extraction)
  → Pass 2: Sonnet 4.6 (replication briefs, top ads only)
  → PostgreSQL `competitor_ads` (multi-tenant via tenant_id)
  → Rating Page `/compete/[tenantSlug]` (Hebrew UI, dark glassmorphism)
  → Client Feedback → RAG Update → Content Strategy

Apify Actor: curious_coder/facebook-ads-library-scraper — pay-per-result, $0.75/1K ads, 18K users, 4.7 rating. No monthly subscription.