This changelog highlights releases that materially changed Facebook Ads Library Scraper for users. Dates use UTC; small internal refactors and routine maintenance are intentionally omitted.
Bug fix
- More Instagram profiles now find their advertiser. Handles such as
mamaearth.in, vahdamindia or itsprolicious used to fail with "No ads were searched or charged" even though the brand advertises. The Actor now also searches the handle's words and the profile's display name, and still uses only the advertiser whose linked Instagram account is exactly that handle. Profiles with no linked advertiser still end without a charge.
Bug fix
- Fixed runs that failed with "Dataset writes failed" when the Apify platform was slow to confirm a save. The Actor now checks whether the ads were saved and saves them again only if they weren't, so you get every ad once and pay once.
- Fixed some advertiser Page searches that failed with "All attempted search tasks failed" when Facebook briefly throttled the first page of results. The Actor now continues on a fresh connection.
- Inputs from other Ad Library tools now work:
searchTerms, keywords and query are read as searchQueries, viewAllPageId as a Page ID, and single pageUrl and adsLibraryUrl values as URLs. These runs used to fail with "Please provide at least one search keyword".
Bug fix
- Searches with fewer ads than
maxAds finish faster. When Facebook marks the first results page as its last, the Actor now returns those ads straight away instead of retrying through a browser. Same ads and same charges as before.
- Fixed runs that occasionally stopped at the start and timed out without scraping or charging anything. When a request to the Apify platform stalls, the Actor now retries it after 30 seconds instead of waiting until the run timed out.
Bug fix
- Many Facebook Page URLs that failed with "No ads were searched or charged" now work, because Pages are now also identified through the Ad Library instead of only the Page itself, which Facebook hides from logged-out visitors for most smaller Pages.
/p/Name-ID and /Name-ID links, including non-Latin names, are also accepted. A Page we can't identify with certainty still fails uncharged, and its message suggests pasting the Ads Library URL.
Bug fix
- Runs started with a short timeout (under 100 seconds, for example
timeoutSecs: 60) now search instead of stopping before their first task. If the timeout is too short to work at all, the error says so.
New features
- Ad details: the new
includeAdDetails option adds an adDetails object to each ad with Meta's "See ad details" data: EU total reach, reach by country, age and gender, targeted locations, ages and gender, payer and beneficiary, and the advertiser's Instagram username, followers and verification. Same price per ad; off by default.
- Most recent ranking:
sortBy: "recent" uses Meta's "Most recent" ranking, which favours recently launched ads (not strict start-date order). A sort in a pasted Ads Library URL is now kept on every page.
- Exact phrase:
searchType: "keyword_exact_phrase" sends Meta's exact-phrase search (Meta may match text that is not in the exported fields).
- New fields
collationId, pageIsDeleted and isCatalogAd.
Bug fix
- Catalog (dynamic product) ads no longer output
{{product.name}}-style placeholders in bodyText, title, linkDescription or linkCaption. The first product card's real text is used, or null when there is none.
- Every run now ends with a status message: how many ads were saved, empty searches, ad-details coverage, or the input error.
Bug fix
- Advertiser Page URLs now return only that Page's ads. Named Pages (
facebook.com/nike), profile.php?id= and /people/ links are resolved to the Page's Ads Library ID; previously a named Page ran a name search and returned, and charged for, ads from other advertisers.
- Non-Facebook URLs, non-Page Facebook links (groups, posts) and Pages that cannot be found are skipped with a clear status message and are never searched or charged. A run with only such inputs fails before any charge.
- Page lookups now count toward the run's time limit, and the scraper also respects a run timeout shorter than the default, so a slow lookup cannot push a run past its timeout before results are saved and charged.
Reliability
- Searches where Facebook reports results but shows none now finish successfully with an empty result instead of failing. When the first results page is empty but more pages exist, the actor follows them and returns the ads Facebook shows later.
- Accepts
urls, startUrls, adLibraryUrls, pageIds and searchQuery from other tools. Ads Library links work with or without a trailing slash. Single-ad links (?id=) get a clear message.
- A results page that reports an error is never treated as "no ads"; the run retries and recovers instead.
- A stalled proxy connection can no longer hold a run past its time limit. Large batches stop on time and keep the ads already found instead of timing out.
Reliability
- Terminal input, extraction and dataset-write errors correctly mark the run as failed instead of reporting success. Acknowledged dataset chunks are retained and charged even if another chunk fails; rejected writes are not automatically retried.
- Searches and pages with no ads in the Ads Library finish in seconds with an empty result instead of retrying.
- The first results page loads over plain HTTP on every run, with automatic browser recovery. If Facebook throttles pagination, the run continues on a fresh connection instead of stopping early. Ad fields, limits and charging are unchanged.
- Keep ads already retrieved over HTTP if a later route reports conflicting empty results or browser recovery is exhausted or runs out of time. Date filters, deduplication and per-ad charging remain unchanged.
Reliability
- First-page retrieval can use HTTP with automatic browser recovery; repeated records do not count toward the requested unique-ad limit. Ad fields, limits and charging are unchanged.
- Run logs now summarize extraction work and recovery paths. Failed operations send sanitized diagnostics to our private error service for repair, as described in the README.
Reliability
- Large multi-page batches now use shorter bounded result-detection windows and preserve enough run time for clean shutdown instead of being killed at the timeout.
Reliability
- Multi-search runs now process two searches at a time and stop prolonged pagination throttling sooner, reducing timeout failures while preserving ads already extracted.
Fixes
- Combined searches now deduplicate overlapping ads by Facebook Ad Library ID across the entire run, preventing duplicate dataset rows and per-ad charges.
- Added duplicate-skip totals to run logs for clearer billing and output diagnostics.
Features
- Added opt-in inclusive ad activity date-window filters with absolute dates or relative durations.
- Date-filtered searches now keep paginating until the requested number of matching ads is collected or results are exhausted.
Features
- Refreshed public Store task examples to better cover common competitor ad monitoring workflows.
Reliability
- Improved extraction fallback behavior so searches can return ads when residential proxy exits time out.
Features
- Added public Store task examples so users can start common Facebook Ads Library workflows faster.
Changes
- Added workload caps and clearer progress logging for long ad searches.
Features
- Added support for pasting Facebook Ads Library results URLs directly as scraper input.
- Clarified setup around no-login scraping: users do not need Meta app review or API tokens.
Reliability
- Reduced proxy cost and bandwidth use with smarter waiting and tighter blocking behavior.
Changes
- Reworked the Actor to use a hybrid browser and HTTP approach for more practical Ads Library extraction.
Fixes
- Improved pagination by using direct GraphQL calls instead of relying only on scrolling.
- Added residential proxy handling and category support for broader Ads Library coverage.
Features
- Initial release for scraping public Meta/Facebook Ads Library results, including ad creatives, copy, links, media, platforms, spend information when available, and Page metadata.