Google Ads Scraper: Competitor Ad Monitor (Transparency Center)

See every Google ad your competitors run (Search, YouTube, Display, Shopping), straight from Google's own public Ads Transparency Center, and get a change feed: which ads are new, which stopped, which restarted since your last run.

Not affiliated with, endorsed or sponsored by Google. Google Ads and Ads Transparency Center are Google trademarks, used only to say where the public data comes from.

What it does

• Enter competitor websites (hubspot.com), brand names (HubSpot) or Google advertiser IDs (AR10072600183532683265). A brand name matches up to 3 Google advertisers whose name is that brand (plus company words like Inc or a country code), not ones that merely contain it.
• Pick a region (US, GB, CA, AU, DE, … or ANYWHERE).
• Get one row per ad: advertiser, format (text / image / video), first and last date shown, days shown, running or stopped, a link to the ad on Google's Transparency Center, and (for text and image ads) a link to Google's archived preview image.
• Schedule it daily or weekly with Only new, stopped and restarted ads on, and you get a clean competitor-ad alert feed. The first run is always the full baseline.

What it does NOT do (honestly)

• No spend, impressions, clicks or keywords: Google does not publish those for commercial ads.
• No targeting data (Google only shows it for ads served in the EU; not included yet).
• Does not download or copy ad images or videos. You get links to Google's own pages; the creatives belong to the advertisers.
• Skips political and election ads (advertisers carrying Google's election-ads verification, or named like a campaign or PAC).
• Skips advertisers that may be private individuals. Google's data doesn't say person vs company, so an advertiser is kept only when its name has a company word (Inc, LLC, Ltd, GmbH…) or matches the website it advertises. A name like "John Smith" is left out and counted in the run summary.
• Advertiser-ID inputs whose name has no company word are skipped for the same reason; enter the website domain instead.
• "Stopped" means Google's own last shown date is more than 3 days old. An ad missing from one run is never marked stopped (a capped or failed search proves nothing).

Output (one row per ad)

change (baseline / new / unchanged / stopped / restarted), status (running / stopped), target, advertiserId, advertiserName, adDomain, creativeId, format, firstShown, lastShown, daysShown, region, adUrl, imageUrl, checkedAt. Run summary (key-value store SUMMARY): per competitor, ads found, Google's own estimate, running, new, stopped, skipped political, skipped individual, errors.

Rules this tool follows (checked 2026-10-13)

• robots.txt: adstransparency.google.com/robots.txt returns 404 (no rules). The tool re-checks every run and stops on purpose if Google ever adds rules that block its data calls.
• Google Terms of Service (effective 2026-07-30): forbid automated access in violation of machine-readable instructions such as robots.txt, bypassing protective measures, and using content to violate others' rights. This tool: no login, no Google account, no CAPTCHA or block bypassing, about 1 request per 1.2 s with backoff on 429/5xx, facts and links only (no ad files copied).
• The Transparency Center is a public registry Google runs under ad-disclosure rules (e.g. EU Digital Services Act, Art. 39).
• Data source: the same public data calls the Transparency Center web page makes. Google can change them at any time; the daily self-test fails loudly when that happens.

Self-test (daily, run by Agent HQ)

node test/selftest.js --save checks 3 real advertisers (hubspot.com, mailchimp.com, canva.com) in the US plus a brand-name search (Mailchimp): rows complete, no emails or phone numbers, links only, second run reports only changes, a known election advertiser is skipped, a known person-named advertiser is skipped, charges counted. Last result: selftest/last.json.

Input example

{
  "targets": [
    "hubspot.com",
    "Mailchimp"
  ],
  "region": "US",
  "onlyChanges": true,
  "maxAdsPerTarget": 200
}

Output example (one row, real data from a test run on 2026-10-13)

{
  "change": "baseline",
  "target": "hubspot.com",
  "advertiserId": "AR10072600183532683265",
  "advertiserName": "Hubspot, Inc.",
  "adDomain": "hubspot.com",
  "creativeId": "CR06039634602782359553",
  "format": "text",
  "firstShown": "2025-07-16",
  "lastShown": "2026-10-13",
  "daysShown": 456,
  "status": "running",
  "region": "US",
  "adUrl": "https://adstransparency.google.com/advertiser/AR10072600183532683265/creative/CR06039634602782359553?region=US",
  "imageUrl": "https://tpc.googlesyndication.com/archive/simgad/1769084597065357731",
  "checkedAt": "2026-10-13T15:17:00.046Z"
}

Pricing draft (pay per event)

Run start $0.005 · each ad row $0.0015 ($1.50 per 1,000). Michael sets the final price.

Folder

• src/core.js pure logic (parsing, skip rules, change detection)
• src/main.js fetching, pacing, saving
• src/runtime.js Apify on the platform, plain files on the Mac
• test/ offline tests + live self-test
• .actor/ Apify config (input, output, Docker)