Google Ads Transparency Scraper (cached)

Every ad an advertiser is running, straight from Google's Ads Transparency Center — creative id, format, preview URL, and the first/last date each ad was shown. Search advertisers by name, or go straight to an AR… id. No API key, no browser.

What you get

Advertiser rows (from keywords):

Field	Example
advertiserName	NIKE SRL
advertiserId	AR17365672681860497409
advertiserCountry	IT
declaredAdCount	17 — Google's own count
advertiserUrl	link to the Transparency Center page

Ad rows (one per creative):

Field	Example
creativeId	CR18300947740944760833
format	image or rendered_preview
previewUrl	direct image URL, or Google's rendered-preview URL
imageHtml	the <img> snippet Google returns, for image ads
firstShownAt, lastShownAt	ISO 8601
creativeUrl	link to the ad in the Transparency Center
fromCache	whether this came from cache or a live fetch

How it behaves

Measured against the live endpoints on 2026-08-06:

• Zero ads is a real answer, not a failure. Google keeps advertisers whose ads have stopped running and still reports a count for them — 3 of 20 advertisers tested returned no creatives for that reason. Those come back as an explicit row with adCount: 0 and declaredAdCount beside it, so you can see the difference. A scraper that returned silence here would look broken.
• An advertiser that fails does not fail the run. It comes back as a row with error, so a 20-advertiser job returns the other 19.
• Pagination stops when the advertiser runs out, not when a page counter says so. Asking for 100 ads from an advertiser with 9 returns 9 and spends nothing extra.
• Google rate-limits this endpoint per IP — about 25 requests from a single address was enough to earn a 429 in testing. So this Actor paces itself, caches every result, and rotates to a fresh proxy IP instead of sleeping through a limit. Two runs asking for the same advertiser within the cache window cost one request to Google, not two.

Limits — read before buying

• region is Google's internal enum, not a country code, and a wrong value returns zero ads with no error. Leave it empty (the default) to get ads from all regions. This Actor will not guess a value for you, because guessing would silently look like "this advertiser has no ads".
• format is reported from the ad content actually present. formatCode passes Google's raw code through untouched — the full enum is not published, and inventing labels for unseen values would be a guess dressed up as data.
• The Transparency Center covers ads Google chooses to disclose. It is not a complete record of an advertiser's spend, and it carries no impression or budget figures.
• Ads Transparency has no official public API and Google's Terms of Service restrict automated access. You are responsible for how you use the output.
• No personal data is collected — advertiser and creative records only.

Input

{
  "keywords": ["nike"],
  "advertisersPerKeyword": 5,
  "maxAdsPerAdvertiser": 100
}

Already know the advertiser? Skip the lookup:

{
  "advertisers": ["https://adstransparency.google.com/advertiser/AR13536189912721653761"],
  "maxAdsPerAdvertiser": 40
}

Only need the ids, not the ads (much cheaper):

{ "keywords": ["nike", "adidas"], "fetchAds": false }

Development

pip install -r requirements.txt
mkdir -p storage/key_value_stores/default
echo '{"keywords":["nike"],"useApifyProxy":false}' > storage/key_value_stores/default/INPUT.json
python -m src

src/ads.py is plain standard library and can be exercised without the Apify SDK — useful when checking whether Google changed the request shape:

$python -c "from src.ads import AdsClient; c=AdsClient(); print(c.search_advertisers('nike')[:3])"