Facebook Ads Library Scraper & Creative Intelligence

Turn a competitor website into a dataset of active Facebook and Instagram ads with copy, headlines, CTAs, creative media, landing URLs, start dates and ad age. Get a ready-to-use intelligence report highlighting longest-running ads, repeated creatives, likely winner candidates, offers, hooks and creative angles.

Built for agencies, ecommerce teams and marketers who need a one-time creative research export. No Meta login, cookies or AI key required.

Quick start

{"domain":"allbirds.com","maxAds":100,"country":"ALL"}

1. Enter a competitor domain.
2. Run the Actor.
3. Export the default dataset as JSON, CSV or Excel.
4. Open Creative intelligence report in the Output tab for the ranked summary.

What you get

Each dataset row is one unique, explicitly active ad from a Facebook page linked on the competitor homepage, or an ad with a verified landing URL on the domain when homepage discovery is unavailable:

Field	Meaning
adId, pageId, pageName	Ad and advertiser identity
adCopy, headline, cta	Public ad text; variants are preserved in separate arrays
creative, imageUrls, videoUrls	Original image and video links when available
creativeVariants	Individual carousel card text, CTA, media and destination
landingUrl, landingUrls	Public destination URLs; Facebook link wrappers are decoded
startDate, adAgeDays	Public start date and elapsed calendar duration
adLibraryUrl	Link back to the original Meta Ad Library record
intelligence	Duration, repetition, winner-candidate, offer, hook and angle evidence
missingFields	Fields not provided by the source

The INTELLIGENCE key-value record contains the longest-running ranking, repeated creative groups, likely winner shortlist, source run IDs, filtering counts, coverage limitations and a link to the companion tracker. Unknown fields remain null or empty; they are never invented.

How discovery works

First, the Actor reads the competitor's HTTPS homepage and finds its Facebook page links. It scrapes active ads from those pages, including ads that open Facebook collections rather than the brand website. Each row includes attributionMethod: HOMEPAGE_FACEBOOK_LINK and the homepage and Facebook page URLs as evidence.

If the homepage is unavailable or has no usable Facebook link, the Actor searches Meta Ad Library for the domain text, verifies destination hostnames, then expands active ads from discovered advertiser pages. In this fallback mode it filters all results by destination domain, using attributionMethod: LANDING_DOMAIN. For example, allbirds.com.kw is not a match for allbirds.com. All results are deduplicated by ad ID.

A link is attribution evidence, not legal proof of brand ownership. Homepage links can identify regional or partner pages; fallback domain searches can include affiliates advertising your website. In fallback mode, ads without an identifiable matching destination, including unresolved shortened links, are excluded.

Set maxAds to 0 to request all accessible ads on discovered pages. Discovery caps, source budgets, Meta search behavior and availability still apply. The report always sets completeCoverageGuaranteed: false; it never claims a complete census. An empty result means no verified matches in this search, not proof of no advertising.

Intelligence methodology

• Longest Running Ads: active ads sorted by public start-date age. This is elapsed age, not verified uninterrupted delivery time.
• Repeated Creatives: same advertiser and same exact set of media URL paths, ignoring expiring query parameters. Distinct ad IDs are counted once. Missing media never creates a repetition group. This is not perceptual image similarity.
• Likely Winners: candidates must be at least longRunningDays old and reuse media across at least two returned ads. Score: min(70, round(ageDays / threshold * 35)) + min(30, (repetitions - 1) * 10). Reasons are included. Score is a heuristic, not a probability.
• Offer: explicit English phrases such as discounts, free shipping, trials, BOGO and limited-time offers, with exact evidence.
• Hook: the opening sentence or line of ad copy, up to 280 characters.
• Creative Angle: English text cues for offer, social proof, convenience, product benefit, sustainability or problem/solution. Unrecognized angles remain empty.

No spend, conversion, ROAS or profitability data is inferred. No OCR, video transcription or visual content analysis is performed. Rankings and repetition counts describe the returned sample.

Input options

Input	Default	Notes
domain	Required	Public website domain or URL
country	ALL	Uppercase two-letter country code or ALL
maxAds	100	0–10,000; 0 requests all accessible results
longRunningDays	30	7–365 days
discoveryLimit	200	30–1,000 keyword search results
maxPages	5	1–20 homepage-linked or discovered advertiser pages
maxSourceChargeUsd	2	$0.10–$50 combined child scraping budget

Pricing

$2 per 1,000 analyzed ads ($0.002 for each output ad), plus a $0.00005 Actor start event at the default memory setting. No analysis charge for an empty dataset or failed collection.

Source scraping charges are additional. This Actor calls Facebook Ads Library Scraper by curious_coder using your Apify execution context. Its current price is shown on its own pricing page. Discovery records, nonmatching ads and repeated reads can incur source charges even when excluded from the final dataset. maxSourceChargeUsd caps the aggregate child-run budget; the parent run's maximum charge controls this Actor's fees separately. Failed source runs may still incur source fees.

For example, 100 analyzed results cost $0.20 in analysis fees plus the small start fee and whatever the source runs actually consumed. The report records source run IDs and accounted source charges for auditing. Because final billing aggregates can settle after a child finishes, budget accounting uses the run's event prices as a conservative floor.

Track changes after scraping

Use Facebook Ads Library Competitor Tracker to track new ads, stopped ads and creative changes over time. This scraper does not maintain a monitoring baseline. The intelligence report includes suggested Facebook page URLs for the tracker; select a specific country supported by that Actor.

Reliability and limitations

Source errors fail the run rather than masquerading as an empty successful scrape. Source scraping can change when Meta changes its interfaces. Public dates and media are returned as provided; CDN URLs may expire. Dynamic template placeholders are excluded when real card copy is available. Missing dates are never treated as zero-day ads. Use the Issues tab with your run ID and input domain if a run fails.

This independent tool is not affiliated with Meta. It processes public ad-library content and does not access private advertiser accounts.