Meta Ad Library Page Resolver

Resolves a brand/company name to candidate Meta/Facebook advertiser pages from Meta Ad Library typeahead.

This actor is intended to replace scrapeio~facebook-ad-library-suggestions with a narrower, cheaper, and more auditable resolver. It does not scrape ads. It only resolves candidate advertiser pages so the app can decide which page ID to pass to the Meta ads scraper.

Input

{
  "brand_name": "Apple",
  "website_domain": "apple.com",
  "country": "ALL",
  "limit": 10
}

Output

Dataset rows use the same core contract as the current actor:

{
  "adlibrary_url": "6629105669",
  "brand_name": "Apple",
  "category": "Product/service",
  "avatar_url": "https://...",
  "likes": 16100000,
  "verification": "BLUE_VERIFIED",
  "page_alias": "apple",
  "ig_username": "apple",
  "ig_followers": 36600000,
  "ig_verification": true,
  "profile_url": "https://www.facebook.com/apple",
  "score": 95,
  "source": "typeahead"
}

Current Implementation Notes

The HAR captures showed the relevant endpoint:

POST https://www.facebook.com/api/graphql/
fb_api_req_friendly_name=useAdLibraryTypeaheadSuggestionDataSourceQuery
doc_id=9755915494515334
variables={
  "queryString": "Apple",
  "isMobile": false,
  "country": "ALL",
  "adType": "ALL"
}

The response already contains the page ID and all useful identity fields under:

data.ad_library_main.typeahead_suggestions.page_results[]

The actor uses impit for browser-impersonated direct HTTP requests and does not launch a browser.

Recommended architecture:

• Actor: extraction and normalization only.
• Adside app: final matching with deterministic rules and LLM fallback when candidates are ambiguous.

Country Behavior Observed

For Apple, ALL, FR, and US return effectively the same top candidates:

1. Apple TV
2. Apple
3. Apple Music / other Apple-adjacent pages

The country parameter is still real and present in the GraphQL variables. Keep it configurable, but prefer ALL for global company resolution unless a workspace has an explicit market.