Meta Ad Library Scraper — Facebook & Instagram Ads + EU Reach

Every ad a page is running on Facebook, Instagram, Messenger, WhatsApp, Audience Network and Threads — as one flat row per creative, with the copy, the landing URL, the media, the run dates and, when you ask, the EU reach. Headline, body, description, call to action, image and video URLs, first and last shown, platforms, and for ads about social issues, elections or politics the spend, impressions and reach ranges Meta publishes with the funding entity. Watch a competitor on a schedule and get only their new ads — into Slack or your inbox through Apify's own integrations, with ads you already have never returned or charged. Same column names as our Google Ads Transparency Scraper, so one sheet holds both platforms. Proxies are handled for you.

Why this scraper, not the other Ad Library actors

• Every ad, not one per creative. The Ad Library lists one ad for each group that shares a creative and text ("N ads use this creative") and counts them all: Shopify shows ~5,700 results behind ~1,750 groups. Most actors in this category page the list and stop at what it shows — the two biggest, curious_coder and igolaizola, open every group and deliver 5,597 (measured 2026-09-20, both, identical sets). This actor opens every group too (expandVariants, on by default), walks an advertiser in the order that does not repeat itself, and continues past the source's cursor limit in media and platform slices. Switch expandVariants off for the creative-level view: one row per group with collation_count.
• Typed columns for the figures that matter, not a nested node to unpick. Ten actors already flatten Meta's record; the three most-used hand back the raw GraphQL node. What none of them types: the EU reach as eu_total_reach, the country × age × gender breakdown as a list, the audience as eu_audience_age_min … eu_audience_locations, the payer and beneficiary; the political ranges as spend_min/spend_max, impressions_min/impressions_max, reach_min/reach_max with the currency and paid_for_by. Carousel and dynamic creatives keep every card in cards and put the first one's copy in the top-level columns, so a text column is never blank when the creative has words somewhere.
• What the ad says and shows, in the same row. transcribeVideos turns the speech into transcript, its language, the first three seconds (hook_3s) and speech_seconds — a music-only video is checked and not billed. ocrImageAds reads the words baked into the artwork with a vision model. analyzeAds returns a typed read: hook, offer, angle, audience, CTA intent, tone, sentiment. enrichLandingPage follows the link to the page it lands on and reads its product name, price and currency. In this category those are four separate actors (steadyfetch, dltik, apify's e-commerce event, automation-lab's resolver), each with its own bill.
• Monitoring that tells you what ended. onlyNewAds skips every creative earlier runs delivered — not returned, not charged. Every row carries first_seen_at and last_seen_at from the same memory, and includeEndedAds adds an ended row for each creative a complete scan no longer finds. The memory is named after the targets, or after your stateStoreName.
• Honest counts, and the limits named. The status of every run says what was delivered next to what the page claimed, and why a walk stopped: the source's cursor limit (about 1,900 creatives in one order, measured), a filter that excluded everything, a spending cap. Nothing is presented as complete that is not.
• A run that stops can be continued. A checkpoint is written after every page. Resurrect a timed-out run, or pass its id as resumeFromRunId, and it carries on from the cursor with no creative delivered or billed twice and maxAds counted across both.
• Files that outlive Meta's links. image_url and video_url expire in about four days. downloadMedia copies the files into a key-value store you own, keyed by creative, and links them from media_files; a creative met again is linked to its existing file, not fetched or billed twice.
• No start fee, and the proxies are ours. Twenty-three actors in this category charge a start fee; this one bills per row only. The Ad Library limits datacenter IPs on their first request and each residential exit after a few hundred; this actor carries its own rotating residential exits — a walker per target and a pool for the details and variant calls — and moves to a fresh one the moment the source pushes back, resuming the same cursor.
• Fast where it counts. 500 Shopify ads with details in 103 s, against 122 s for curious_coder and 297 s for igolaizola on the same input the same afternoon (2026-09-20); three advertisers walked in parallel, each on its own exit.
• The same columns as our Google Ads Transparency Scraper. advertiser_id, creative_id, headline, body_text, destination_url, cta, first_shown, last_shown, days_shown, image_url, spend_min, region, ad_url mean the same thing in both. Schedule both against a competitor and one sheet holds their Google and Meta advertising side by side.

Who it's for

• Performance marketers & agencies — read a competitor's live creatives — the copy, the offer, the landing page, the formats they are testing — and how long each has been running.
• Creative strategists & copywriters — a corpus of ad copy that is proven in market, with run dates showing which messages survived, instead of a swipe file nobody knows performed.
• Founders sizing up a market — who is paying to acquire customers in a niche, on which platforms, with what positioning — evidence of a market with money in it, before you spend any.
• Researchers, journalists & compliance teams — political and issue advertising with spend, impressions, reach and the funding entity, and the EU transparency figures Meta publishes per ad.

Quick start

Every active ad one page runs

{
  "advertiserIds": [
    "20409006880"
  ],
  "region": "ALL",
  "maxAds": 100
}

With EU reach and audience per ad

{
  "advertiserIds": [
    "20409006880"
  ],
  "includeEuReach": true,
  "maxAds": 50
}

Political ads about a topic, with spend

{
  "searchTerms": [
    "election"
  ],
  "adType": "political_and_issue",
  "region": "US",
  "maxAds": 50
}

Monitoring — only what appeared since the last run

{
  "advertiserIds": [
    "20409006880"
  ],
  "onlyNewAds": true
}

Output

One record per ad creative:

field	meaning
row_type	ad for a creative; advertiser for a page resolved from advertiserNames / searchTerms in resultType: advertisers; count for a claimed-total row in resultType: counts; advertiser_profile and landing_page for the rows summaryRows adds; ended for a creative a complete scan no longer returns (includeEndedAds).
advertiser_name	The page's name as the Ad Library shows it.
advertiser_id	The numeric Facebook page ID behind the ads. Join on this across runs and platforms; it is also what advertiserIds takes.
creative_id	The ad's Library ID — the id in an Ad Library URL. Unique per creative; the key for onlyNewAds.
collation_id	Meta groups near-identical creatives (the same ad run in several sizes or placements) under one collation. The group's id.
collation_count	How many ads share this creative and text — Meta's group ("N ads use this creative"). With expandVariants on, every one of them is a row and they share the collation_id; off, this row stands for the group.
ad_format	image, video, dco (a dynamic creative with per-card copy and media), carousel, multi_images and the other display formats Meta uses, lowercased.
is_active	Whether the ad was running when the run read it.
headline	The ad's title. For carousel and dynamic creatives, the first card's title when the top level has none.
body_text	The primary text. For carousel and dynamic creatives, the first card's body when the top level has none.
link_description	The description line under the link, when the creative carries one.
headlines	Every card's title, for carousel and dynamic creatives. Empty on single-card ads.
descriptions	Every card's body text, for carousel and dynamic creatives.
display_url	The caption shown under the creative — usually the advertiser's domain.
destination_url	Where the ad sends a click.
cta	The call-to-action button's label, e.g. Sign up, Shop now.
cta_type	Meta's own CTA code, e.g. SIGN_UP, SHOP_NOW.
cards	For carousel and dynamic creatives, one entry per card: headline, body_text, link_description, display_url, destination_url, cta, cta_type, image_url, video_url, video_preview_url.
first_shown	The day the ad started running (UTC).
last_shown	The day the ad was last seen running — today for an active ad.
days_shown	Days between first_shown and last_shown, inclusive. Long-running ads are the ones that earn their budget.
image_url	The creative's image — the ad's own, or the first card's. A direct CDN link that fetches without a login and expires about four days after the run.
image_urls	Every distinct image URL the creative carries, in order.
image_small_url	Meta's ~600 px rendition of the first image — the one ocrImageAds reads; image_url is the original.
video_url	The creative's video, HD rendition where there is one. Same expiry as the images.
video_urls	Every distinct video of the creative, in order (multi-video and carousel creatives carry several).
video_duration_s	Length of video_url in whole seconds, read from the URL's own encode tag (efg.duration_s, present on 80 of 80 video ads measured 2026-09-20) — no extra request. Null for image ads.
video_preview_url	The poster frame of the video.
media_files	With downloadMedia: one entry per file copied into your key-value store — kind (image/video), key, url (the store URL, which does not expire), bytes, content_type, source_url, and reused (true when an earlier run already stored it — linked, not billed again). Empty otherwise.
image_file_url	With downloadMedia: the store URL of the first image. Null otherwise.
video_file_url	With downloadMedia: all: the store URL of the first video. Null otherwise.
transcript	With transcribeVideos: what is said in the video, as one text. Null when the video has no speech (then speech_seconds is 0) or when it was not transcribed.
transcript_language	With transcribeVideos: the spoken language as an ISO 639-1 code (en, fr, ko), as detected.
hook_3s	With transcribeVideos: the words spoken in the first three seconds of speech — the hook — counted from the first word, not from the start of the video.
speech_seconds	With transcribeVideos: seconds of speech in the video. 0 means the video was checked and carries no speech (music or silence); null means it was not transcribed.
ad_ocr_text	With ocrImageAds: the words a vision model read out of the creative's image (or a video's poster frame), in reading order. Null when the image carries no readable text or was not read.
sentiment_score	On every row, free: the mean AFINN-165 valence of the English words in the copy, scaled to -1 (negative) .. 1 (positive). Null when the copy has no scored word — a non-English ad is unscored, not neutral.
analysis_hook	With analyzeAds: the line meant to stop the scroll, quoted or closely paraphrased from the ad.
analysis_offer	With analyzeAds: the concrete thing on offer — product, discount, trial, free item — or null when the ad makes no specific offer.
analysis_angle	With analyzeAds: the persuasion angle — problem_solution, social_proof, scarcity, discount, feature, lifestyle, curiosity, education, comparison, brand or other.
analysis_audience	With analyzeAds: who the ad is written for, in the model's words (e.g. small retailers opening a second location).
analysis_cta_intent	With analyzeAds: what the ad wants done — purchase, signup, download, learn_more, contact, visit, watch, apply, book or other.
analysis_tone	With analyzeAds: playful, urgent, professional, friendly, luxurious, informative, emotional, bold or other.
analysis_sentiment	With analyzeAds: the model's read of the copy's emotional valence, -1 .. 1.
landing_final_url	With enrichLandingPage: the page the ad's link lands on after redirects, with tracking parameters (utm_*, fbclid, gclid …) stripped. Null when nothing answered.
landing_domain	With enrichLandingPage: the final page's domain, without www..
landing_status	With enrichLandingPage: the HTTP status the final page answered (200, 404, 403 …); null when the site did not answer at all.
landing_title	With enrichLandingPage: the page's title (Open Graph title first, then <title>).
landing_is_product	With enrichLandingPage: true when the page declares itself a product (JSON-LD Product, og:type product, or a price in its markup); false for other pages; null when the page did not answer.
landing_product_name	With enrichLandingPage: the product's name from JSON-LD, or the page title on a product page.
landing_price	With enrichLandingPage: the price the page publishes for machines (JSON-LD offers, product:price:amount), as shown to the run's location — a storefront that localises by visitor answers in that country's currency. Null when the page publishes none.
landing_currency	With enrichLandingPage: the ISO currency of landing_price.
platforms	Where the ad ran: facebook, instagram, messenger, whatsapp, audience_network, threads.
launched_last_7_days	Whether the ad started running within the last 7 days of the run.
launched_last_30_days	Whether the ad started running within the last 30 days of the run.
long_running	Whether the ad has been shown for at least longRunningDays (default 30) — the creatives that survive are the ones that convert.
platform_count	How many Meta platforms the ad ran on.
categories	Meta's categories for the ad — political for ads about social issues, elections or politics; unknown for ordinary commercial ads.
regions	Countries the ad targeted or reached, when Meta lists them (political ads).
region	The country the run asked for — a code, ALL, or a comma-separated list when regions held several.
spend_min	Lower bound of the spend range Meta publishes. Political and issue ads only; null on commercial ads.
spend_max	Upper bound of the spend range; null when the range is open-ended (>$1M).
spend_currency	ISO currency of the spend range.
impressions_min	Lower bound of the impressions range. Political and issue ads only.
impressions_max	Upper bound of the impressions range; null when open-ended (>1M).
reach_min	Lower bound of the estimated reach range. Political and issue ads only.
reach_max	Upper bound of the estimated reach range; null when open-ended.
paid_for_by	The disclaimer's funding entity on a political or issue ad.
funding_entity_website	The funding entity's website, from Meta's verification record.
funding_entity_address	The funding entity's address, from Meta's verification record.
eu_total_reach	Accounts reached across the EU, as published under the DSA. Filled by includeEuReach on ads that reached the EU; null on the rest.
eu_reach_breakdown	Reach per EU country, age range and gender: entries of country, age_range, male, female, unknown. Filled by includeEuReach.
eu_audience_age_min	The youngest age the advertiser targeted in the EU.
eu_audience_age_max	The oldest age the advertiser targeted in the EU.
eu_audience_gender	The gender the advertiser targeted in the EU: All, Men or Women.
eu_audience_locations	The locations the advertiser included or excluded in the EU: entries of name, excluded, type.
payer	Who paid for the ad, as disclosed under the DSA. Filled by includeEuReach.
beneficiary	Who benefits from the ad, as disclosed under the DSA.
political_reach_by_region	For political ads with includeEuReach: the share of reach per region, entries of region and reach_share (0–1), regions with reach only.
political_reach_by_age_gender	For political ads with includeEuReach: the share of reach per age range and gender.
advertiser_category	The page's category, e.g. Software. Filled by includeAdvertiserInfo or includeEuReach.
advertiser_about	The page's About text. Filled by includeAdvertiserInfo or includeEuReach.
advertiser_categories	The page's categories as the Ad Library lists them, e.g. ["Software"].
advertiser_ads_seen	On an advertiser row: how many ads of the name search belonged to this page.
advertiser_match	On an advertiser row: how the page name matched the name searched — exact, starts, contains or none.
advertiser_claimed_ads	On an ad row from a page target: the total the Ad Library page claims for that advertiser — ads, not creative groups. With expandVariants on, the delivered count comes close; the status says what the source's cursor limit cut (see the FAQ).
claimed_count	On a count row: the total the Ad Library shows for the target, from one page load.
summary_ad_count	On a summary row: how many delivered ads it covers.
summary_active_count	On an advertiser_profile row: how many of them are active.
summary_longest_days	On an advertiser_profile row: the longest days_shown among them.
summary_formats	On an advertiser_profile row: {format, count} per ad format, most common first.
summary_landing_domains	Distinct landing-page domains behind the ads.
summary_distinct_copies	On an advertiser_profile row: how many distinct primary texts the ads use.
summary_ad_ids	On a summary row: the Library IDs of the ads it covers.
advertiser_likes	The page's like count.
advertiser_ig_username	The page's linked Instagram handle. Filled by includeAdvertiserInfo or includeEuReach.
advertiser_ig_followers	Followers of the linked Instagram account.
advertiser_verified	Whether the page carries Meta's verification badge.
advertiser_spend_total	For political pages: lifetime spend across the page's disclaimers, as Meta reports it. Filled by includeAdvertiserInfo or includeEuReach.
advertiser_spend_week	For political pages: spend in the last seven days.
advertiser_url	The page's Facebook URL.
advertiser_picture_url	The page's profile picture.
ad_url	The ad in the Ad Library, for a human to open.
matched_query	The keyword or Library ID that produced this row; null on rows from a page ID.
matched_fields	On a keyword row: which of the ad's texts contain the whole search term — headline, body_text, link_description, destination_url, display_url, advertiser_name, cards. Empty means the Library matched it loosely; keywordRelevance: visible drops those.
first_seen_at	With a monitoring memory: the day a run of these targets first saw the creative.
last_seen_at	With a monitoring memory: the day a run last saw it — today on a live row.
ended_at	On an ended row: the day a complete scan first failed to return the creative.
run_tag	The runTag you gave the run, on every row of every type — so rows from several runs sort in one sheet. Null when none was given.
raw	Meta's own record, untouched — only with includeRawData.

Example record:

{
  "row_type": null,
  "advertiser_name": "Shopify",
  "advertiser_id": "20409006880",
  "creative_id": "1657481092030542",
  "collation_id": "1308187987545965",
  "collation_count": 1,
  "ad_format": "dco",
  "is_active": true,
  "headline": "{{product.name}}",
  "body_text": "{{product.brand}}",
  "link_description": "{{product.description}}",
  "headlines": [
    "Vendez gratuitement sur Shopify",
    "Jetzt kostenlos auf Shopify verkaufen"
  ],
  "descriptions": [
    "Obtenez le bouton que tout le monde adore. Commencez à vendre et à gagner de l’argent avec Shopify 💰",
    "Der Button, den die Leute kennen und lieben. Jetzt verkaufen und mit Shopify verdienen 💰"
  ],
  "display_url": "shopify.com",
  "destination_url": "https://www.shopify.com/se/gratis-trial",
  "cta": "Sign up",
  "cta_type": "SIGN_UP",
  "cards": [
    {
      "headline": "Vendez gratuitement sur Shopify",
      "body_text": "Obtenez le bouton que tout le monde adore. Commencez à vendre et à gagner de l’argent avec Shopify 💰",
      "link_description": null,
      "display_url": "shopify.com",
      "destination_url": "https://www.shopify.com/fr/essai-gratuit",
      "cta": "Sign Up",
      "cta_type": "SIGN_UP",
      "image_url": null,
      "video_url": "https://video-mxp2-1.xx.fbcdn.net/o1/v/t2/f2/m366/AQOMRBFZd7QAA0ruNoYgN6JXjAiwMNc_Yicgl9hZIh8lOlORAPEtR50aqEB4yfQkm6GjRg_8HrdX7-RYLqyVQCzvuyhfySfpOGVKhVn39EzgmQ.mp4?_nc_cat=101&_nc_sid=b66105&_nc_ht=video-mxp2-1.xx.fbcdn.net&_nc_ohc=dEUkWdMHDwYQ7kNvwHiJ8XN&efg=eyJ2ZW5jb2RlX3RhZyI6Inhwdl9wcm9ncmVzc2l2ZS5WSV9VU0VDQVNFX1BST0RVQ1RfVFlQRS4uQzMuNzIwLmRhc2hfaDI2NC1iYXNpYy1nZW4yXzcyMHAiLCJ4cHZfYXNzZXRfaWQiOjgxNTc1MTk4NzU2ODUxNSwiYXNzZXRfYWdlX2RheXMiOjM1OSwidmlfdXNlY2FzZV9pZCI6MTAxMzksImR1cmF0aW9uX3MiOjEyLCJ1cmxnZW5fc291cmNlIjoid3d3In0%3D&ccb=17-1&vs=257c57522b2c310e&_nc_vs=HBksFQIYRWZiX2VwaGVtZXJhbC9CODQzMDVENjk4MkMzOThBMjBFRTg0MzRBMkUyRDg5N19tdF8xX3ZpZGVvX2Rhc2hpbml0Lm1wNBUAAsgBEgAVAhhAZmJfcGVybWFuZW50LzExNEI1QjA2OTQzNEMyQzVCMjVGODRERUJCQjJFQjhEX2F1ZGlvX2Rhc2hpbml0Lm1wNBUCAsgBEgAoABgAGwKIB3VzZV9vaWwBMRJwcm9ncmVzc2l2ZV9yZWNpcGUBMRUAACaG_oz3gfvyAhUCKAJDMywXQCgAAAAAAAAYGWRhc2hfaDI2NC1iYXNpYy1nZW4yXzcyMHARAHUAZbaeAQA&_nc_gid=CXhKjE0I-jAPsnY2HWAufw&_nc_ss=7e289&_nc_zt=28&oh=00_AQKP-fMbUXalHsS9wXs3LMR-_48s_pNZB8-Q4gsPXfn6Nw&oe=6AB52807",
      "video_preview_url": "https://scontent-mxp1-1.xx.fbcdn.net/v/t39.35426-6/710447423_2716333165389931_221950439945272582_n.jpg?_nc_cat=106&ccb=1-7&_nc_sid=c53f8f&_nc_ohc=Iq_OjuwglZMQ7kNvwFbTh7K&_nc_oc=Adpm8fPUfnukFoj49quE74I_yce8YnDZH9bV7imwtVDPiUtTJFYVW9EdKEwXNHklqYY&_nc_zt=14&_nc_ht=scontent-mxp1-1.xx&_nc_gid=CXhKjE0I-jAPsnY2HWAufw&_nc_ss=7e289&oh=00_AQI43eq2bcH5RMsal09q_k9dBmEqZ4ebUy7tpCOHT7Yc9Q&oe=6AB51813"
    },
    {
      "headline": "Jetzt kostenlos auf Shopify verkaufen",
      "body_text": "Der Button, den die Leute kennen und lieben. Jetzt verkaufen und mit Shopify verdienen 💰",
      "link_description": null,
      "display_url": "shopify.com",
      "destination_url": "https://www.shopify.com/de/kostenloser-test",
      "cta": "Sign Up",
      "cta_type": "SIGN_UP",
      "image_url": null,
      "video_url": "https://video-mxp1-1.xx.fbcdn.net/o1/v/t2/f2/m366/AQMzymmwQZ2QnkOpNdfkNUDVtUuEF_DCKrHfcjYtpPp0SVWyHsjV3IHAZNshaq4BvtK4tZKod6N2t8ZH1me2ehX7_coU5YYzbFp7Fqetyh3Fng.mp4?_nc_cat=107&_nc_sid=b66105&_nc_ht=video-mxp1-1.xx.fbcdn.net&_nc_ohc=ZQvNX5y0_1gQ7kNvwFWdBMd&efg=eyJ2ZW5jb2RlX3RhZyI6Inhwdl9wcm9ncmVzc2l2ZS5WSV9VU0VDQVNFX1BST0RVQ1RfVFlQRS4uQzMuNzIwLmRhc2hfaDI2NC1iYXNpYy1nZW4yXzcyMHAiLCJ4cHZfYXNzZXRfaWQiOjEyNTgyNDg1NDI3NDA4NjUsImFzc2V0X2FnZV9kYXlzIjozNTksInZpX3VzZWNhc2VfaWQiOjEwMTM5LCJkdXJhdGlvbl9zIjoxMiwidXJsZ2VuX3NvdXJjZSI6Ind3dyJ9&ccb=17-1&vs=70346b3d005c775a&_nc_vs=HBksFQIYRWZiX2VwaGVtZXJhbC9BMTQ4Mzk4RkE3M0IwMkVBNDgzNjRCODA5MDZFNjBCMV9tdF8xX3ZpZGVvX2Rhc2hpbml0Lm1wNBUAAsgBEgAVAhhAZmJfcGVybWFuZW50LzlGNDlFMUVFMDhFNzFEQzAyNEUzQzIyMDI4RDcxOTkxX2F1ZGlvX2Rhc2hpbml0Lm1wNBUCAsgBEgAoABgAGwKIB3VzZV9vaWwBMRJwcm9ncmVzc2l2ZV9yZWNpcGUBMRUAACaChu-W2pe8BBUCKAJDMywXQCgAAAAAAAAYGWRhc2hfaDI2NC1iYXNpYy1nZW4yXzcyMHARAHUAZbaeAQA&_nc_gid=CXhKjE0I-jAPsnY2HWAufw&_nc_ss=7e289&_nc_zt=28&oh=00_AQLzjp0zQpdcDwTblT9m2eHv5WFYDx4hLzEUSMmU_TpnjA&oe=6AB5139D",
      "video_preview_url": "https://scontent-mxp1-1.xx.fbcdn.net/v/t39.35426-6/711695694_2404378190031495_3175246454924053399_n.jpg?_nc_cat=102&ccb=1-7&_nc_sid=c53f8f&_nc_ohc=8n96Gr_0IrEQ7kNvwHWtXgQ&_nc_oc=AdqVInXxe25_YfyFpmh849bsCEOj61j5_c_9sTORV1Zd9ZUVJu5bB12kx7TKh-PcXzo&_nc_zt=14&_nc_ht=scontent-mxp1-1.xx&_nc_gid=CXhKjE0I-jAPsnY2HWAufw&_nc_ss=7e289&oh=00_AQKyNhZtD7LZhNeGGIbt0hnRLCgoPYbTXKH6-TCNdN5L9w&oe=6AB5359C"
    }
  ],
  "first_shown": "2026-06-03",
  "last_shown": "2026-09-19",
  "days_shown": 109,
  "image_url": null,
  "image_urls": [],
  "image_small_url": null,
  "video_url": "https://video-mxp2-1.xx.fbcdn.net/o1/v/t2/f2/m366/AQOMRBFZd7QAA0ruNoYgN6JXjAiwMNc_Yicgl9hZIh8lOlORAPEtR50aqEB4yfQkm6GjRg_8HrdX7-RYLqyVQCzvuyhfySfpOGVKhVn39EzgmQ.mp4?_nc_cat=101&_nc_sid=b66105&_nc_ht=video-mxp2-1.xx.fbcdn.net&_nc_ohc=dEUkWdMHDwYQ7kNvwHiJ8XN&efg=eyJ2ZW5jb2RlX3RhZyI6Inhwdl9wcm9ncmVzc2l2ZS5WSV9VU0VDQVNFX1BST0RVQ1RfVFlQRS4uQzMuNzIwLmRhc2hfaDI2NC1iYXNpYy1nZW4yXzcyMHAiLCJ4cHZfYXNzZXRfaWQiOjgxNTc1MTk4NzU2ODUxNSwiYXNzZXRfYWdlX2RheXMiOjM1OSwidmlfdXNlY2FzZV9pZCI6MTAxMzksImR1cmF0aW9uX3MiOjEyLCJ1cmxnZW5fc291cmNlIjoid3d3In0%3D&ccb=17-1&vs=257c57522b2c310e&_nc_vs=HBksFQIYRWZiX2VwaGVtZXJhbC9CODQzMDVENjk4MkMzOThBMjBFRTg0MzRBMkUyRDg5N19tdF8xX3ZpZGVvX2Rhc2hpbml0Lm1wNBUAAsgBEgAVAhhAZmJfcGVybWFuZW50LzExNEI1QjA2OTQzNEMyQzVCMjVGODRERUJCQjJFQjhEX2F1ZGlvX2Rhc2hpbml0Lm1wNBUCAsgBEgAoABgAGwKIB3VzZV9vaWwBMRJwcm9ncmVzc2l2ZV9yZWNpcGUBMRUAACaG_oz3gfvyAhUCKAJDMywXQCgAAAAAAAAYGWRhc2hfaDI2NC1iYXNpYy1nZW4yXzcyMHARAHUAZbaeAQA&_nc_gid=CXhKjE0I-jAPsnY2HWAufw&_nc_ss=7e289&_nc_zt=28&oh=00_AQKP-fMbUXalHsS9wXs3LMR-_48s_pNZB8-Q4gsPXfn6Nw&oe=6AB52807",
  "video_urls": null,
  "video_duration_s": null,
  "video_preview_url": "https://scontent-mxp1-1.xx.fbcdn.net/v/t39.35426-6/710447423_2716333165389931_221950439945272582_n.jpg?_nc_cat=106&ccb=1-7&_nc_sid=c53f8f&_nc_ohc=Iq_OjuwglZMQ7kNvwFbTh7K&_nc_oc=Adpm8fPUfnukFoj49quE74I_yce8YnDZH9bV7imwtVDPiUtTJFYVW9EdKEwXNHklqYY&_nc_zt=14&_nc_ht=scontent-mxp1-1.xx&_nc_gid=CXhKjE0I-jAPsnY2HWAufw&_nc_ss=7e289&oh=00_AQI43eq2bcH5RMsal09q_k9dBmEqZ4ebUy7tpCOHT7Yc9Q&oe=6AB51813",
  "media_files": null,
  "image_file_url": null,
  "video_file_url": null,
  "transcript": null,
  "transcript_language": null,
  "hook_3s": null,
  "speech_seconds": null,
  "ad_ocr_text": null,
  "sentiment_score": null,
  "analysis_hook": null,
  "analysis_offer": null,
  "analysis_angle": null,
  "analysis_audience": null,
  "analysis_cta_intent": null,
  "analysis_tone": null,
  "analysis_sentiment": null,
  "landing_final_url": null,
  "landing_domain": null,
  "landing_status": null,
  "landing_title": null,
  "landing_is_product": null,
  "landing_product_name": null,
  "landing_price": null,
  "landing_currency": null,
  "platforms": [
    "facebook",
    "instagram",
    "messenger"
  ],
  "launched_last_7_days": null,
  "launched_last_30_days": null,
  "long_running": null,
  "platform_count": null,
  "categories": [
    "unknown"
  ],
  "regions": [],
  "region": "ALL",
  "spend_min": null,
  "spend_max": null,
  "spend_currency": null,
  "impressions_min": null,
  "impressions_max": null,
  "reach_min": null,
  "reach_max": null,
  "paid_for_by": null,
  "funding_entity_website": null,
  "funding_entity_address": null,
  "eu_total_reach": 11900089,
  "eu_reach_breakdown": [
    {
      "country": "HU",
      "age_range": "55-64",
      "male": 9965,
      "female": 9552,
      "unknown": 165
    },
    {
      "country": "HU",
      "age_range": "45-54",
      "male": 28878,
      "female": 15165,
      "unknown": 539
    },
    {
      "country": "HU",
      "age_range": "Unknown",
      "male": 85,
      "female": 15,
      "unknown": 3
    }
  ],
  "eu_audience_age_min": 18,
  "eu_audience_age_max": 65,
  "eu_audience_gender": "All",
  "eu_audience_locations": [
    {
      "name": "Guadeloupe",
      "excluded": false,
      "type": "countries"
    },
    {
      "name": "Aland Islands",
      "excluded": false,
      "type": "countries"
    }
  ],
  "payer": "Shopify",
  "beneficiary": "Shopify",
  "political_reach_by_region": [],
  "political_reach_by_age_gender": [],
  "advertiser_category": "Software",
  "advertiser_about": null,
  "advertiser_categories": null,
  "advertiser_ads_seen": null,
  "advertiser_match": null,
  "advertiser_claimed_ads": null,
  "claimed_count": null,
  "summary_ad_count": null,
  "summary_active_count": null,
  "summary_longest_days": null,
  "summary_formats": null,
  "summary_landing_domains": null,
  "summary_distinct_copies": null,
  "summary_ad_ids": null,
  "advertiser_likes": 4538277,
  "advertiser_ig_username": "shopify",
  "advertiser_ig_followers": 2670355,
  "advertiser_verified": true,
  "advertiser_spend_total": null,
  "advertiser_spend_week": null,
  "advertiser_url": "https://www.facebook.com/shopify/",
  "advertiser_picture_url": "https://scontent.fqrc1-1.fna.fbcdn.net/v/t39.30808-1/532226807_1206379304854792_3907785642549358122_n.jpg?stp=dst-jpg_s148x148_tt6&_nc_cat=1&ccb=1-7&_nc_sid=418b77&_nc_ohc=yojCiHRPRY4Q7kNvwGR-8_b&_nc_oc=Adp5zqArc6xx9hsnNw2-EYHw4eLMMZ49xkrwMX29M0tmUmkCHREDl7smamLE02Cskg4&_nc_zt=24&_nc_ht=scontent.fqrc1-1.fna&_nc_gid=seJgawSzPhkdwcP5u-9nbA&_nc_ss=7e289&oh=00_AQKgCRxis0r2Mjv2KY7tp0qyvceiUdbG2025-XHTVxgm5A&oe=6AB52A47",
  "ad_url": "https://www.facebook.com/ads/library/?id=1657481092030542",
  "matched_query": null,
  "matched_fields": null,
  "first_seen_at": null,
  "last_seen_at": null,
  "ended_at": null,
  "run_tag": null,
  "raw": null
}

Pricing

Pay only for what a run actually returns, with no start fee. Prices are per 1,000 and fall with your Apify plan:

per 1,000	Free	Starter	Scale	Business
Ad creative — copy, link, CTA, media URLs, run dates, platforms	$1.00	$0.85	$0.75	$0.65
EU reach and audience, when the ad reached the EU (includeEuReach)	+$0.80	+$0.70	+$0.60	+$0.50
Political spend, impressions and reach, when Meta published them	+$1.00	+$0.85	+$0.75	+$0.65
Advertiser (page) resolved from a name (resultType: advertisers)	$2.00	$1.70	$1.50	$1.30
Count row (resultType: counts)	$10.00	$8.50	$7.50	$6.50
Creative file copied into your store (downloadMedia), per file written	$4.00	$3.50	$3.00	$2.50
Video transcript, when the video has speech (transcribeVideos)	+$30.00	+$26.00	+$23.00	+$20.00
Image text read by a vision model, when words were found (ocrImageAds), model usage included	+$1.00	+$1.00	+$1.00	+$1.00
Typed analysis — hook, offer, angle, audience, CTA intent, tone, sentiment (analyzeAds), model usage included	+$10.00	+$8.50	+$7.50	+$6.50
Landing page resolved and read, when the page answered (enrichLandingPage)	+$5.00	+$4.50	+$4.00	+$3.50

Each fee is charged per row that actually carries the thing. A commercial ad is never billed the political fee; an ad that never reached the EU is never billed the EU fee; a repeat the source served twice is never returned, so never billed; a file an earlier run already stored is linked, not billed again; a music-only video is checked and not billed. includeAdvertiserInfo is one request per advertiser and is not billed separately. Enterprise plans have their own rates — the Actor's Pricing tab shows the price for your plan.

Usage patterns

• Follow a competitor weekly — Put their page ID in advertiserIds, tick onlyNewAds, save it as a task and schedule it. Each run returns only creatives the earlier runs did not — the first run delivers the current set, every later run only what appeared since. Connect the task to Slack or email through Apify's integrations and the new ads land where you read.
• Read what they dropped — activeStatus: inactive returns the ads a page has stopped running. Next to the active set, with first_shown and last_shown, it is the history of what they tried and for how long.
• Compare against their Google ads — Run the same competitor through our Google Ads Transparency Scraper. The two outputs share advertiser_name, headline, body_text, destination_url, cta, first_shown, last_shown and days_shown, so one sheet shows both platforms.
• Read what the creatives say — transcribeVideos, ocrImageAds and analyzeAds together turn a competitor's feed into a table you can pivot: what each video says in its first three seconds, what the artwork promises, and a typed read — angle, offer, audience, CTA intent, tone — per ad. Add enrichLandingPage for the page each ad sells, with its price.
• Measure a political campaign — adType: political_and_issue with a keyword or a page ID, region set to the country, and includeEuReach for the demographic and regional reach shares. Spend, impressions and reach arrive as the ranges Meta publishes.
• Size a competitor before you scrape — resultType: counts with page IDs or brand names returns one row per page with the total the Ad Library shows for it (claimed_count) — from a single page load, no ads scraped, a cent per row. Run it first to see whether a competitor has 60 ads or 6,000, or schedule it weekly and watch the number move. Ready-made: Count how many Meta ads each competitor runs.
• Keep the media — image_url and video_url are direct links to Meta's CDN and fetch without a login; they expire about four days after the run. Set downloadMedia: all and the files are copied into a key-value store of yours, linked from media_files, image_file_url and video_file_url — no expiry, and a creative met again on a later run is linked, not fetched twice.

Input configuration

field	type	default	what it does
advertiserIds	array	[]	Numeric Facebook page IDs, e.g. 20409006880 (Shopify) — the view_all_page_id value in an advertiser's Ad Library URL. The whole Ad Library URL is accepted too. Every ad the page runs, across Facebook, Instagram, Messenger, WhatsApp, Audience Network and Threads.
advertiserNames	array	[]	Brand or company names, e.g. Nike. Each is resolved to the Facebook pages advertising under that name — the pages behind the ads a name search returns, ranked by how well the page name matches — and those pages are then scraped (resultType: ads) or listed (resultType: advertisers). Use maxAdvertisersPerName to control how many pages per name are followed.
searchTerms	array	[]	Keywords or brand names, e.g. shopify. A keyword search is Meta's ranked sample across every advertiser — the Ad Library returns at most a few hundred creatives for any keyword, whatever total it displays. To get everything one advertiser runs, use its page ID instead. keywordMatch chooses loose or exact-phrase matching.
creativeIds	array	[]	Library IDs of individual ads, e.g. 1657481092030542 — the id in an Ad Library URL (the whole URL is accepted too). One lookup each.
resultType	ads / advertisers / counts	"ads"	ads returns one record per ad creative. advertisers returns one record per page that advertiserNames or searchTerms resolve to — page id, name, categories, likes, how many matching ads it runs — so you can see which page a brand advertises under before pulling its ads (the same idiom as our Google Ads Transparency Scraper). counts returns one record per target with the total the Ad Library shows for it, from a single page load and without scraping any ad.
maxAdvertisersPerName	integer	3	How many resolved pages to follow for each entry in advertiserNames. A common name can match several pages (regional pages, resellers); the best name match ranks first.
startUrls	array	[]	Paste Ad Library URLs — an advertiser page (...?view_all_page_id=...), a keyword search (...?q=...) or a single ad (...?id=...). A Facebook page URL (facebook.com/shopify) works too: the Actor finds the advertiser whose page name is that name, and returns nothing rather than a different page when no name matches. A country in the URL is used unless you set region explicitly.
region	string		Two-letter country code for where the ads were shown, e.g. US, GB (UK also works), DE, or ALL for every country at once — the same country= value the Ad Library puts in its URLs. Leave it empty and the country in a pasted startUrls URL is used, or ALL when there is none. The EU transparency figures (includeEuReach) do not depend on this: they come back for any ad that reached the EU whatever country you ask for.
regions	array	[]	Several country codes in one run, e.g. ["DE", "FR", "IT"] — ads shown in any of them. Wins over region when given; ALL in the list means every country. Rows carry the list in region.
targetedCountryOnly	boolean	false	With a country in region/regions: ask the Ad Library for ads whose advertiser targeted that country, not every ad that happened to be shown there — its own 'targeted' switch, passed through as is. Measured 2026-09-20 on a page and on a keyword search in DE: the first 30 rows were identical with and without it, so expect a small difference at most. No effect with ALL.
activeStatus	active / inactive / all	"active"	active returns ads running now, inactive ads that have stopped, all both. Inactive ads are how you read a competitor's history: what they tried and dropped.
adType	all / political_and_issue / housing / employment / financial_products	"all"	all for every ad. political_and_issue narrows to ads about social issues, elections or politics — the only ads Meta publishes spend, impressions and reach for; those arrive as ranges in spend_min/spend_max, impressions_min/impressions_max and reach_min/reach_max, with the funding entity in paid_for_by. housing, employment and financial_products are Meta's special ad categories.
mediaType	all / image / video / meme / image_and_meme / none	"all"	Keep only creatives of this media type, as the Ad Library's own filter defines them. all is every type.
platforms	array	[]	Keep only ads that ran on these Meta platforms. Empty means all. Every row also lists where it ran in platforms.
languages	array	[]	Keep only creatives in these languages, as two-letter ISO 639-1 codes, e.g. ["es", "pt"] — the Ad Library's own language filter. Empty means all.
sortBy	auto / relevance / most_recent / most_impressions / least_impressions	"auto"	auto picks the order that reaches the most: for an advertiser the newest-first order walks every creative in a straight line (measured: relevance starts repeating after ~1,240), for a search term relevance yields more. most_impressions / least_impressions order by Meta's impressions bucket, which exists on ads about social issues, elections or politics only.
expandVariants	boolean	true	The Library groups ads that share a creative and text ("N ads use this creative") and shows one per group; its result count counts them all. On, each group is opened and every ad in it is a row of its own — the count the Library shows, the way the lane's biggest actors deliver it (Shopify: 1,753 creatives, 5,597 ads). Off, one row per creative with collation_count saying how many ads use it: fewer rows, fewer fees, the creative-level view.
keywordMatch	unordered / exact_phrase	"unordered"	How searchTerms match ad text: unordered finds ads containing the words in any order, exact_phrase only the phrase as typed. A pasted search URL keeps its own search_type.
keywordRelevance	any / visible	"any"	A keyword search is the Ad Library's loose match: shopify also returns ads that mention the word in passing. Every keyword row carries matched_fields — which of its texts (headline, body, cards, link, advertiser name) contain the whole term. visible keeps only rows with at least one such field; any keeps everything the Library returned.
dedupeBy	none / advertiser	"none"	advertiser keeps one ad per advertiser in a keyword search — a who-is-advertising-on-this read instead of every creative. none keeps all.
startDate	string		YYYY-MM-DD. Keeps ads whose run overlaps this date or later — an ad that started years ago and is still running counts, because it is still running.
endDate	string		YYYY-MM-DD. Keeps ads whose run overlaps this date or earlier.
period	`` / last24h / last7d / last14d / last30d / last90d	""	Keep only creatives that started running in the window: the last 24 hours, 7, 14, 30 or 90 days. A preset over first_shown; combine with sortBy: most_recent for a 'what did they launch this week' read. Empty means no window.
longRunningDays	integer	30	An ad shown for at least this many days gets long_running: true — the fastest read on which creatives earn their budget. Default 30.
includeEuReach	boolean	false	One extra request per ad that adds what the Ad Library publishes under the EU Digital Services Act: total EU reach, the reach breakdown by country, age and gender, the audience the advertiser chose (age range, gender, included and excluded locations) and the payer and beneficiary. Present on ads that reached the EU, null on the rest; billed only on rows that carry a figure. Also fills the advertiser columns below.
includeAdvertiserInfo	boolean	false	One extra request per advertiser (not per ad) that fills advertiser_category, advertiser_about, advertiser_likes, advertiser_ig_username, advertiser_ig_followers, advertiser_verified and, for political pages, lifetime and last-week spend. Not billed separately.
downloadMedia	none / images / all	"none"	Copy each creative's files from Meta's CDN into a key-value store you own, so the links do not expire with Meta's ~4-day URLs. Every row then carries media_files (store URL, size, type per file) plus image_file_url / video_file_url. Files are keyed by creative id, so a creative met again on a later run is linked to its existing file and not fetched or charged twice. Billed per file written.
mediaStoreName	string		Named key-value store the files go into. Optional: left empty, the name is derived from the targets (meta-ads-media-…), so a scheduled task keeps adding to the same store. Set one to collect several inputs into one place, e.g. creatives-shopify.
transcribeVideos	boolean	false	Turn the speech in every video creative into text: transcript, transcript_language (ISO code), hook_3s (the words spoken in the first three seconds of speech) and speech_seconds. A music-only or silent video is checked and comes back with speech_seconds: 0 — and is not billed. Billed per row that carries a transcript.
ocrImageAds	boolean	false	Read the words baked into the artwork with a vision model — the headline set over the photo, the offer sticker, the price — into ad_ocr_text. The image read is the creative's first image, or the poster frame of a video ad. Read once per creative and cached. Billed per row where words were found; model usage is included.
analyzeAds	boolean	false	One structured read per creative by an OpenAI model: analysis_hook (the scroll-stopper), analysis_offer, analysis_angle (problem_solution, social_proof, scarcity, discount, feature, lifestyle, curiosity, education, comparison, brand, other), analysis_audience, analysis_cta_intent, analysis_tone and analysis_sentiment (-1..1). Reads the copy, the cards, the CTA, and the image text and transcript when those switches are on. A free sentiment_score (AFINN word list) is on every row regardless. Billed per row analysed; model usage included.
enrichLandingPage	boolean	false	Follow each ad's link to the page it lands on: landing_final_url (redirects followed, tracking parameters stripped), landing_domain, landing_status, landing_title, and — from the page's own product markup (Open Graph, JSON-LD) — landing_is_product, landing_product_name, landing_price, landing_currency. One fetch per distinct URL per run. Billed per row whose page answered.
openaiApiKey	string		Optional. The Actor uses its own model access by default and bills a flat fee per transcript, image read or analysis. Supply your own OpenAI key only if you would rather the model usage went to your account.
summaryRows	boolean	false	After the ads, add one advertiser_profile row per advertiser (ad count, active count, longest-running days, formats, landing domains, distinct copies, the ad ids) and one landing_page row per destination URL with the ads pointing at it. Computed from the rows already delivered — no extra requests, not billed.
runTag	string		Your own label for this run — a client name, a campaign, a week — copied into run_tag on every row, so rows from several runs sort in one sheet or table.
maxAds	integer	0	Stop after this many ads in total, across every search. 0 means no limit.
maxAdsPerTarget	integer	0	Cap each page ID, keyword or URL separately. Without this, one prolific advertiser can use the whole maxAds budget before the next target starts. 0 means no per-target cap.
onlyNewAds	boolean	false	Skip creatives delivered by previous runs of the same targets (or of the same stateStoreName). Skipped ads are never returned and never charged — a weekly monitoring run pays only for what actually changed. Every row also carries first_seen_at / last_seen_at from the same memory.
includeEndedAds	boolean	false	With a monitoring memory (the same store onlyNewAds uses): after a complete scan of a target, add one ended row for every creative earlier runs saw there that the Library no longer returns — with when it was first and last seen. A capped or cut-off scan reports none, because it cannot vouch for an absence. Not billed.
stateStoreName	string		Name of the memory that holds the creative IDs already delivered, so onlyNewAds can tell new from old across runs. The memory is kept by the Actor on its side, one per Apify account — nothing is stored in your account. Optional: left empty, the name is derived from the targets, so a scheduled task remembers on its own. Set one to share memory between different inputs, e.g. ads-shopify, or change it to start over.
resumeFromRunId	string		The id of an earlier run of this Actor that timed out or was aborted, with the same input: this run continues from its checkpoint — finished targets skipped, the target in progress continued from its cursor, no creative delivered twice, maxAds counted across both. A run that stops on its own can also simply be resurrected (Apify's Resurrect): it finds its own checkpoint.
requestDelaySecs	number	0.3	Seconds to wait between result pages. The source paces itself by latency; a small delay keeps one exit useful for longer.
concurrency	integer	3	How many advertisers or search terms are paged at the same time, each on its own residential exit (the details and variant calls run on a separate pool of three). 3 is measured safe; 5 for a long list of small targets.
includeRawData	boolean	false	Add Meta's own untouched record to each row as raw, for callers who would rather map the fields themselves.

FAQ

The Ad Library says a page has ~5,700 ads. Will I get them all?

Nearly. The count is ads; the list shows one per creative group, and this actor opens every group (expandVariants, on by default). On Shopify, measured 2026-09-20: 2,310 creatives opened into 5,337 ads, 95% of the 5,597 the two most-used actors deliver, in 19 minutes. The source refuses cursors past roughly 1,900 creatives in any one order; when a walk hits that, the actor continues in media-type and platform slices and the status names what was cut. A keyword search is different: it is Meta's ranked sample, a few hundred creatives whatever the count says, on every actor.

Why are spend_min, impressions_max and reach_min null on my ads?

Meta publishes spend, impressions and reach only for ads about social issues, elections or politics. Commercial ads never carry them, on the Ad Library site or anywhere else — every actor in this category returns null there, whatever its schema promises. Use adType: political_and_issue for the ads that have them. For commercial ads the one reach figure that exists is the EU one: switch on includeEuReach.

What does includeEuReach cost me?

One extra request per ad, and the per-row EU fee only on rows that came back with a figure — an ad that never reached the EU is not billed for it. includeAdvertiserInfo alone is one request per advertiser and is not billed separately.

Can I search by a brand name instead of a page ID?

Yes — put it in searchTerms. A keyword search is Meta's ranked sample across every advertiser mentioning the words, which is right for a topic and wrong for a page: it caps at a few hundred creatives and mixes advertisers. To get everything one page runs, open it in the Ad Library and copy the view_all_page_id from the URL into advertiserIds, or paste the URL into startUrls.

Which region should I use?

ALL returns every ad the page runs anywhere. A country code narrows to ads shown there. The EU figures do not depend on it: includeEuReach returns them for any ad that reached the EU, whatever country you asked for.

The image or video link stopped working.

Meta's CDN links carry an expiry of about four days from the moment the run fetched them. Download what you need within that window; a new run issues fresh links.

What does a PARTIAL status mean?

The source refused to keep answering even after the actor rotated through fresh exits, or answered with errors on every attempt, for the searches named in the status. Whatever was delivered is in the dataset; the named searches are incomplete. It is rare, and it is ours to fix — a PARTIAL run is a bug report, not a configuration problem on your side.

Integration

JavaScript

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: 'YOUR_TOKEN' });
const run = await client.actor('hyperbach/meta-ads-library-scraper').call({"advertiserIds": ["20409006880"], "region": "ALL", "maxAds": 100});
const { items } = await client.dataset(run.defaultDatasetId).listItems();

Python

from apify_client import ApifyClient
client = ApifyClient('YOUR_TOKEN')
run = client.actor('hyperbach/meta-ads-library-scraper').call(run_input={'advertiserIds': ['20409006880'], 'region': 'ALL', 'maxAds': 100})
items = client.dataset(run['defaultDatasetId']).list_items().items

CLI

$apify call hyperbach/meta-ads-library-scraper --input '{"advertiserIds": ["20409006880"], "region": "ALL", "maxAds": 100}'

REST

curl -X POST "https://api.apify.com/v2/acts/hyperbach~meta-ads-library-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H 'Content-Type: application/json' -d '{"advertiserIds": ["20409006880"], "region": "ALL", "maxAds": 100}'

Support

apify@hyperbach.com

This page is generated from the Actor's schemas and a live sample — it cannot describe a field the Actor does not have.