Google Ads Scraper

Scrape public Google Ads Transparency Center creatives by advertiser ID, destination domain, or search term. Export Google Ads data to JSON, CSV, or Excel for competitive research, creative review, and advertiser checks. This scraper is an alternative to building and maintaining a custom connection to the site's public search interface.

The actor returns individual ad creatives or one row per advertiser. Creative rows include the ad ID, advertiser identity, format, activity dates, days served, direct image or hosted preview link, and a link back to the public record. Optional full details add the available creative variations and country reach. No account on the source site is needed.

What data can you extract from Google Ads?

Choose ads for creative records or advertisers for account summaries. A dataset contains one kind of row per run, selected by resultType. Empty fields are returned as null when the public source does not provide them. The two modes share the public advertiser ID and requested region.

Field	Description	Example
creativeId	Public creative identifier	CR15481450763665801217
advertiserId	Public advertiser identifier	AR16735076323512287233
advertiserName	Name shown with a creative	Available in creative rows
domain	Matched destination domain when exposed	example.com
format	Creative type	IMAGE
firstShown	First recorded activity in UTC	2026-08-25T21:30:48.312Z
lastShown	Last recorded activity in UTC	2026-09-27T04:15:56.459Z
shownForDays	Number of days on which the ad was served	8
previewType	Type of preview returned	image
imageUrl	Direct image asset, when present	Hosted image URL
previewUrl	Hosted preview, when present	Hosted preview URL
width, height	Pixel dimensions in the image markup	348, 200
googleCreativeId	Creative identifier embedded in some previews	String or null
googleCustomerId	Customer identifier embedded in some previews	String or null
adGroupId	Ad group identifier embedded in some previews	String or null
versionId	Version identifier embedded in some previews	String or null
adUrl	Direct link to the public creative page	Transparency Center URL
region	Region selected for the run	US
variationCount	Number of returned variations with full details	3
variations	Preview information for every available variation	Array of objects
regionsShown	Country reach with last shown dates	Array of objects
containsSyntheticMedia	Source indicator shown in full details	false
name	Displayed name in advertiser mode	Public advertiser name
countryCode	Advertiser billing or disclosure country	US
approxAdCountLow	Lower bound of the source's approximate ad count	7000
approxAdCountHigh	Upper bound of the source's approximate ad count	8000
legalName	Disclosed legal name, when available	Public legal name
verified	Verification status reported by the source	true
matchedQuery	Search term or domain that produced the match	example.com
advertiserUrl	Link to the public advertiser page	Transparency Center URL

Activity dates describe what the public service reports. shownForDays is a count of served days; it is not the calendar difference between firstShown and lastShown. An ad could be shown intermittently. Likewise, approximate ad-count bounds are ranges from the source, not an exact census.

How to scrape Google Ads

1. Enter an advertiser ID, an exact destination domain, a search term, or a copied Transparency Center URL. An advertiser ID is the AR... value in a public advertiser URL.
2. Choose a region. US is preselected; ANYWHERE searches without a country restriction. Set resultType to ads or advertisers.
3. Set maxAdsPerSearch and maxItems to the amount of data you need. Add a format, platform, or date window if you want a narrower creative set.
4. Start the actor. Results are written as each row is collected. Download the dataset as JSON, CSV, or Excel, or retrieve it through the platform API.

Start with one exact advertiser ID when you already know the account. Use an exact domain when you want to see creatives from multiple advertiser accounts pointing to the same destination. Use a search term when you do not know the ID; the actor resolves public advertiser suggestions and checks the most prominent matches. Search terms can match similarly named accounts, so inspect advertiserId, advertiserName, and the advertiser page before combining records in a report.

Input

Parameter	Type	Default	Description
queries	String list	Empty	Terms or domains resolved through public suggestions.
domains	String list	Empty	Exact destination domains.
advertiserIds	String list	Sample ID	Public AR... IDs.
startUrls	URL list	Empty	Public advertiser or creative URLs, plus search URLs with domain or query.
resultType	Select	ads	Creative rows or advertiser summaries.
region	Select	US	Supported country code or ANYWHERE.
platforms	Select list	Empty	Search, YouTube, Shopping, Maps, or Play; empty means all.
adFormat	Select	ALL	TEXT, IMAGE, VIDEO, or all formats.
startDate	Date text	Empty	Keep ads whose reported activity overlaps this date or later.
endDate	Date text	Empty	Keep ads whose reported activity overlaps this date or earlier.
maxAdsPerSearch	Integer	100	Maximum ad rows read for each target and selected platform.
maxAdvertisersPerQuery	Integer	5	Maximum suggested advertiser IDs inspected per term; at most 10.
maxItems	Integer	100	Total dataset-row cap; zero removes this total cap.
includeDetails	Boolean	false	Fetch variations and country reach with an extra request per ad.
proxyConfiguration	Object	Platform proxy	Network settings; datacenter access is tried first.

At least one of queries, domains, advertiserIds, or startUrls is required. The prefilled ID demonstrates the input shape; replace it to focus on your own target. When several input lists are supplied, the actor processes advertiser IDs, then domains, then individual creative URLs, deduplicating creatives by ID across results. A total cap can end the run before later targets are reached. For balanced coverage, run targets separately or raise maxItems.

{
  "advertiserIds": ["AR16735076323512287233"],
  "resultType": "ads",
  "region": "US",
  "adFormat": "IMAGE",
  "maxAdsPerSearch": 100,
  "maxItems": 100,
  "includeDetails": false
}

To inspect a destination domain across accounts, use domains instead of advertiserIds. Domain results retain the matching domain in each row when the source returns one. For a compact account check, set resultType to advertisers; the actor looks up each advertiser and obtains the source's approximate count bounds. For an individual creative URL, paste the full public URL into startUrls.

Output example

The following fields come from one public image-creative result in the validated 40-row advertiser run. The advertiser name is omitted from this shortened example; the dataset includes it. Preview IDs are null when they are not present in image markup.

{
  "creativeId": "CR15481450763665801217",
  "advertiserId": "AR16735076323512287233",
  "format": "TEXT",
  "firstShown": "2026-08-25T21:30:48.312Z",
  "lastShown": "2026-09-27T04:15:56.459Z",
  "shownForDays": 8,
  "previewType": "image",
  "imageUrl": "https://tpc.googlesyndication.com/archive/simgad/12219066510995162321",
  "previewUrl": null,
  "width": 348,
  "height": 200,
  "adUrl": "https://adstransparency.google.com/advertiser/AR16735076323512287233/creative/CR15481450763665801217?region=US",
  "region": "US"
}

This example illustrates why the public format and preview representation should be interpreted separately: the source marks this creative as TEXT, while its preview is an image. A preview can also be hosted behind a URL rather than exposed as a direct image. Consumers should use format for the source's classification and previewType, imageUrl, and previewUrl to decide how to display a preview.

In advertiser mode, the actor emits advertiserId, name, countryCode, approxAdCountLow, approxAdCountHigh, legalName, verified, matchedQuery, advertiserUrl, and region. Counts can change between runs as activity changes. matchedQuery is null when the row came from a direct ID rather than a term or domain.

Use cases

• Competitive researchers can compare the creatives, formats, and served-day counts for public advertiser accounts in a given country.
• Creative teams can gather image assets and hosted previews for a review board, then inspect full variations for selected ads.
• Sales operations can check a list of public advertiser IDs and receive compact account rows with approximate ad-count ranges.
• Market analysts can compare public ad activity across destination domains, including domains associated with several advertiser IDs.
• Compliance reviewers can keep a dated record of public creative IDs and URLs for their own review process.
• Reporting teams can append scheduled dataset exports and compare newly observed creative IDs with prior exports.

How much does it cost to scrape Google Ads?

The actor's current result event is $0.0016 per dataset row, or $1.60 per 1,000 results. A 40-row run is about $0.064; 500 rows are about $0.80; 5,000 rows are about $8.00. The actor has no separate result event for full details: turning on includeDetails increases the number of source requests and usually increases run time, while the row count remains the billing unit. The price displayed on the actor's Pricing tab is authoritative if pricing changes later.

The maxItems input is the simplest way to cap delivered rows. maxAdsPerSearch limits the work for each target, while maxAdvertisersPerQuery limits how many accounts a free-text term can expand into. If your run has a maximum charge configured on the platform, the actor stops when the dataset push reports that the charge limit was reached. A run with no matching public ads may finish with zero rows.

Tips for best results

Use the most specific input you have. An exact advertiser ID avoids name collisions. An exact domain is useful when the destination is the important entity, but it can return ads from several accounts and languages. A text query is broader: public suggestions can include similar names, regional accounts, or unrelated near matches. Check the IDs before treating all results as one organization.

Keep includeDetails off for exploratory runs. Search results already include IDs, dates, format, and a preview. Turn details on when the additional variations or country list are actually needed, such as a small final selection for a creative report. Each detailed row requires one more lookup, so the elapsed time grows with the number of rows.

Choose a region deliberately. Country scope affects which ads the public service returns, and the same creative may appear in several country-specific runs. ANYWHERE is useful for discovery, but country runs are easier to compare. Supported codes are US, GB, CA, AU, DE, FR, IN, BR, JP, IT, ES, NL, MX, SG, and ZA. A code outside this list must be changed to a supported value or ANYWHERE.

Platform selections are queried separately and deduplicated by creative ID. When you select several platforms, maxAdsPerSearch applies to each selected platform for each target, while maxItems remains the total output cap. Platform filtering follows the public site's own coverage; older ads may not carry complete platform information. The date fields use the public activity window: an ad is kept if that window overlaps your chosen start and end dates. Dates do not mean the creative ran continuously during the interval.

Limits

This actor reads publicly visible records only. The public source decides which ads and accounts are available, when records are updated, and whether counts are precise. The advertiser count is a lower and upper bound, not an exact number. The source can also return an empty result for a valid ID, domain, or country if no public ads meet the filters.

Preview data varies by creative. Direct image links and dimensions appear when the public preview contains image markup. Text, video, and other renderings may instead provide a hosted preview URL. The identifiers embedded in a hosted preview URL are absent from many image rows. Do not use an empty preview ID to conclude that an ad itself is missing; use creativeId as the public record key.

Full-detail regionsShown reflects only the country reach exposed for that creative. Each entry retains a numeric geoCriteriaId; the actor translates supported country IDs to region codes and leaves other codes as null rather than guessing. Full-detail variations can be empty when the public lookup does not expose them. Some individual creative URLs can also fail to return a detail record, even when search still shows a preview.

The actor paginates using the source's returned page token and stops at input limits. It does not claim to retrieve ads hidden by the source, ads that require an account, or every creative that ever ran. Large, broad domains may produce many matches; increase limits or narrow by region, format, and platform when you need a manageable dataset. Search results can shift while a long run is in progress, so deduplication avoids repeated creative IDs within a run.

Use with scheduled workflows

A scheduled run can export the current public records each day or week. Compare creativeId values across datasets to find newly observed ads, and compare lastShown values to see which records changed. For an advertiser check, compare the approximate count range and public ID rather than treating a small count change as a precise increase. Keep the same input and region across snapshots so the comparison is meaningful.

Example requests for an automated workflow are: “Collect up to 100 public image creatives for this advertiser ID in the US and export the dates and image links,” and “Return advertiser summaries for these public IDs with approximate ad-count ranges.” Supply the IDs or domains directly in the input so the scheduled run is reproducible.

FAQ

Do I need an account on the source site? No. The actor reads the public Transparency Center interface. It does not request private account data.

Is it legal to collect these records? The records are publicly visible, but you are responsible for the applicable source terms, local law, and any privacy requirements. Use public advertiser and creative data responsibly, and avoid using the output to infer private information about individuals.

Do I need to configure a proxy? The default uses the platform proxy. The fetch layer begins with datacenter access and can rotate or escalate if it encounters a block. Leave the default in place unless you have a tested network requirement.

How fast is a run? Speed depends on the number of targets, pages, filters, and detail lookups. A 40-row creative validation completed successfully on the platform. Full details add a lookup for every returned creative and therefore take longer than a preview-only run.

Can I schedule or monitor changes? Yes. Schedule repeated runs and compare creative IDs between datasets. This actor does not maintain a persistent seen-ID list or emit only-new rows itself; maintain that comparison in your own workflow.

Can I export to a spreadsheet? Yes. Download the default dataset as CSV or Excel, or send dataset items through an integration. Nested detail arrays are easiest to inspect in JSON.

Why are some fields empty? The public service does not expose every preview identifier, domain, dimension, or country detail for every creative. Empty optional fields are represented as null and are not estimated.

Are advertiser counts exact? No. The public service provides approximate lower and upper bounds. Use them for broad comparisons, and keep both fields when exporting data.

Integrations

Run the actor manually or on a schedule, retrieve results through the platform API, and connect completed runs to webhooks or workflow services. JSON preserves nested creative variations and country reach; CSV and Excel work well for the flat fields. The default dataset URL is available in the actor output. For reliable joins across runs, store creativeId with each creative row and advertiserId with each account row.

Support

Open an issue on the Issues tab with the input, public URL, region, and the field that appears wrong. We reply within 24 hours and add fields on request.