Google Ads Transparency Scraper - New Competitor Ads Monitor
Pricing
from $2.10 / 1,000 ad returneds
Google Ads Transparency Scraper - New Competitor Ads Monitor
For marketers and agencies watching competitors: only the Google Ads Transparency Center ads that are new since your last run, by domain or advertiser ID, with image URL, first/last shown, countries and who paid. On nike.com a rerun returned 0 of 57 older ads as new (2026-09-24).
Pricing
from $2.10 / 1,000 ad returneds
Rating
0.0
(0)
Developer
NeverEmpty
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
3 days ago
Last modified
Categories
Share
Google Ads Transparency Scraper - New Ads Monitor
Watch your competitors in the Google Ads Transparency Center (adstransparency.google.com) and get only the ads that are new since your last run. Give it a website domain (nike.com) or an advertiser ID (AR18378488041124659201); schedule it daily or hourly; each run returns the ads Google started showing after your watch began and that you have not received yet.
Every ad row has the advertiser name and ID, creative ID, format (text, image, video), first and last shown time, days shown, a direct image URL, Google's preview URL and a link to the ad in the Ads Transparency Center. With ad details on (the default) it also adds every country the ad was shown in with first and last shown dates, who paid for the ad, its topic and the image of every variation.
Unofficial. It reads only what the public Ads Transparency Center shows, without logging in.
What it is for
- New competitor ads, every morning: schedule it once a day for a list of competitor domains; each run returns only the ads that appeared since the last run.
- Ad library export: the first run for a target returns its newest ads (20 by default, up to 1,000) and remembers the rest.
- One country or one format:
regionandformatare filters that Google applies itself. Measured on 2026-09-24: Nike Retail BV has about 80,000–90,000 ads anywhere and 19 in Japan — the region filter returns those 19.
How the "new ads" watch works
- First run for a target: reads Google's list (up to
maxAdsScanned, 40 per request), returns themaxAdsFirstRunads with the most recent first-shown time, and remembers every ad it read. - Every later run: reads only ads Google lists as shown since shortly before the previous check, and returns ads that are not remembered and whose first shown time is after the watch started. When the first run read the whole list for the target, ads first shown up to 2 days before the watch started also count, because Google can list an ad a little after it first runs.
- An ad that was already running when the watch started but was not read in the first run (Google does not list ads newest first, and the order changes between requests) is remembered, not reported as new. Measured on 2026-09-24: a second run 20 seconds after the first found 57 such ads for nike.com and returned 0 of them as new.
- When Google has more ads than
maxAdsScannedfor a target, a free row says so, with Google's own estimate of the total. For very large advertisers, narrow the watch withregionorformat.
Watches are kept per target, region, format and watchName. Use a different watchName for two schedules that watch the same target. Do not let two runs of the same watch overlap in time: both would read the same new ads before either remembers them, and both would return (and charge) them; a free row says when a run could not save what it read.
When one run has a domain and that domain's advertiser ID as two targets, an ad found for both is returned once, for the first target.
Input
| Field | What it does |
|---|---|
targets | Domains, advertiser IDs or Ads Transparency Center links, one per line (up to 50). Empty = nike.com. |
region | Two-letter country code (US, GB, DE, JP …) or anywhere. Empty = anywhere. |
format | all, text, image or video. Empty = all. |
maxAdsFirstRun | Ads to return the first time a target is checked (0–1,000). Empty = 20. |
maxAdsScanned | Ads of Google's list to read per target per run (40–1,000). Empty = 400. |
includeAdDetails | Countries with dates, who paid, topic and variation images for each returned ad. Empty = on. |
watchName | Separate watches for the same target. Optional. |
resetMonitoringState | Forget what the watch remembered and start again with a first check. |
Output
One row per ad (status: "ok", changeType: "new"):
| Column | Meaning |
|---|---|
advertiserName, advertiserId | The advertiser as named in the Ads Transparency Center |
creativeId, adUrl | The ad's ID and its page in the Ads Transparency Center |
format | TEXT, IMAGE or VIDEO |
firstShownAt, lastShownAt | First and last time Google shows for the ad (ISO time, UTC) |
daysShown | Number of days shown, as Google reports it |
daysSinceFirstShown | Added: whole days from first shown to the check |
imageUrl, imageWidth, imageHeight | Direct image of the ad when Google archives one. Many text ads are archived by Google as an image of the ad (headline, text and URL as it appeared). |
previewUrl | Google's own preview script for ads that are rendered, not archived as an image (many video and responsive ads). It is not an MP4 file. |
domain | The domain Google shows for the ad (domain searches) |
isFirstCheck | true on the first run for this watch |
sponsorName | Who paid for the ad, as the ad's page in the Ads Transparency Center names it (with ad details on) |
adTopic | Google's topic for the ad when it gives one |
regionsShown | Every country the ad was shown in: countryCode, firstShownDate, lastShownDate, and impression bounds when Google gives them (mainly EU countries) |
variationsCount, variationImageUrls | Number of variations and the image of each archived variation |
euImpressionsLowerBound, euImpressionsUpperBound | Google's EU impression range when it gives one |
detailsRead | false when the ad's detail page could not be read (its detail columns are empty) |
Rows that are not ads are free and say why: no-new-ads, watch-started, no-ads (Google lists nothing for this target — it gives the same answer for a domain it does not know), partial-scan, blocked, unreadable, bad-input, budget-reached.
Pricing
- Run start: charged once per run in which Google's ad list was read for at least one target. Not charged when Google refused every request.
- Ad returned: charged per ad row. Free rows are never charged.
- A run whose maximum total charge has no room for the start fee plus one ad requests nothing and is charged nothing.
- Ad details are included in the ad price (one extra request per returned ad).
Measured on 2026-09-24 (build 0.1.3)
nike.com, first run: 400 ads read on 10 pages, 20 returned with details, in 22 to 65 seconds.- Same watch 20 seconds later: 0 new ads, 57 older unseen ads remembered and not reported.
AR18378488041124659201with regionJP: 19 ads on 1 page, all 19 returned.- Requests from Apify's servers: 34 of 34 answered through datacenter IPs and 33 of 34 through residential IPs (one connection failure); no check page in 102 requests.
Limits
- Google does not list ads newest first. A watch sees every new ad only when the whole list for the target fits in
maxAdsScanned; otherwise a freepartial-scanrow says so. maxAdsScannedis capped at 1,000 per target per run (25 requests).- A domain Google does not know and a domain with no ads get the same empty answer from Google, so both come back as
no-ads. - If Google answers with its "unusual traffic" check page, this Actor does not solve it or keep asking from other IPs: the target comes back as a free
blockedrow and nothing is remembered, so the next run checks it again. - Ad text is not transcribed: an ad comes as Google's archived image of it (
imageUrl) or as Google's preview script (previewUrl). - Political ads are a separate Google library and are not covered.
Support
Something missing or wrong? Open an issue on the Issues tab with the run ID and the target.