Google Images Scraper — Full-Size Image URLs & Sources

Scrape Google Images results for any search term and country: the full-size image URL (not just the thumbnail), width and height, Google's thumbnail, the image title, the source page URL, domain and site name, file size, dominant color and, when Google shows it, the license (Creative Commons / stock licence URL, provider, creator, credit). Use Google's own filters: size, color, type (clip art, line drawing, GIF), time, usage rights and SafeSearch. Export to JSON, CSV or Excel, or call it from your app via API.

The actor collects image links and metadata; it does not download the images.

What data does it extract?

Field	Example	Notes
position	1	Rank in Google's results for this search (and sub-search, see searchVariant)
title	Eiffel Tower - Wikipedia	Title Google shows under the image
imageUrl	https://upload.wikimedia.org/.../Golden_retriever_stehfoto.jpg	Full-size image on the source site
imageWidth, imageHeight	3000, 2000	Original size in pixels
thumbnailUrl	https://encrypted-tbn0.gstatic.com/images?q=tbn:...	Google's thumbnail
thumbnailWidth, thumbnailHeight	547, 365
sourceUrl	https://commons.wikimedia.org/wiki/File:Golden_retriever_stehfoto.jpg	Page the image is on
sourceDomain	commons.wikimedia.org
sourceName	Wikimedia Commons	Site name Google shows
imageId	D6RW5KcSwZ0-zM	Google's image id (stable; used for deduplication and monitoring)
fileSize	2.1MB	As Google labels it
dominantColor	rgb(208,189,170)	Google's placeholder color
licenseUrl	https://creativecommons.org/licenses/by/3.0	Only when Google shows licence details ("Licensable" badge)
licenseAcquireUrl	https://www.gettyimages.com/detail/photo/...	Where to get / license the image
licenseProvider	Getty Images
creator, credit, copyrightNotice	Luka Magic / 500px, ©Getty Images, credit: ...	Image credit as published in the image metadata
isVideo, videoDuration	true, 1:02	Video thumbnails (TikTok, Instagram, YouTube) Google mixes into the results
searchTerm, searchVariant, page	eiffel tower, as asked, 2	Which search, sub-search and Google page the image came from
country, language, scrapedAt	gb, en, 2026-10-11T22:06:54Z

Related searches and refinement chips Google shows for each term ("Puppy", "Cute", "golden retriever black", ...) are saved for free in the key-value store record RELATED_SEARCHES.

Example output

{
  "position": 122,
  "title": "Eiffel Tower Guided Tours | Paris Insiders Guide",
  "imageUrl": "https://www.parisinsidersguide.com/image-files/eiffel-tower-upshot-dramatic-sky-gyg-1000-2x1.jpg",
  "imageWidth": 1000,
  "imageHeight": 500,
  "thumbnailUrl": "https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSIgOYDcK6lwBDHkE6PRlVWD_7F-BxEiAH06JNKcPjORQ&s=10",
  "thumbnailWidth": 632,
  "thumbnailHeight": 316,
  "sourceUrl": "https://www.parisinsidersguide.com/eiffel-tower-guided-tours.html",
  "sourceDomain": "www.parisinsidersguide.com",
  "sourceName": "Paris Insiders Guide",
  "imageId": "jtz9WGYDG_kPfM",
  "fileSize": "246KB",
  "dominantColor": "rgb(232,194,149)",
  "licenseUrl": null,
  "licenseAcquireUrl": null,
  "licenseProvider": null,
  "creator": null,
  "credit": null,
  "copyrightNotice": null,
  "isVideo": false,
  "videoDuration": null,
  "searchTerm": "eiffel tower",
  "searchVariant": "as asked",
  "page": 2,
  "country": "gb",
  "language": "en",
  "scrapedAt": "2026-10-11T22:07:01Z"
}

How to scrape Google Images

1. Enter one or more Search terms (e.g. golden retriever, modern kitchen interior).
2. Pick the Country (google.com, google.co.uk, google.de, ...) and optionally the language.
3. Optionally set Google's filters: Size (large / medium / icon), Color (black and white, transparent or a color), Type (clip art, line drawing, GIF), Time (past 24 hours / week / month / year), Usage rights (Creative Commons, commercial & other licenses) and SafeSearch.
4. Set Max images and run. You can also paste Google Images URLs into Google Images URLs; their search term, country, language, tbs filters and SafeSearch setting are reused.

Example input:

{
  "queries": ["golden retriever"],
  "country": "us",
  "imageSize": "large",
  "usageRights": "creativeCommons",
  "maxItems": 200
}

How many images can I get per search?

Google's results page shows 50-100 images; the next pages (start=10, 20, ...) add about 100 more each until Google answers with an empty page. In our tests that was 165 images for "golden retriever" (pages 1-3, page 4 empty) and 300 images in 3 pages for "eiffel tower" in the UK. When a search runs dry before Max images, the actor runs sub-searches with the Google filters you left open (large, medium, past year, past month, clip art, line drawing, GIF, black and white, transparent, each color, licences, icon) and removes duplicates by Google's image id; searchVariant tells you which sub-search found an image and position is the rank inside it. Max Google pages per search term (default 10) caps how many Google pages each term may use.

How much does it cost?

You pay $0.0025 per image saved (the Image scraped event) plus Apify platform usage: each Google page is one request through Apify's GOOGLE_SERP proxy (about $0.0025 per page of 50-100 images) and a little compute. For 1,000 images that is about $2.50 in events + ~$0.03-0.05 SERP proxy + ~$0.02 compute and storage = ~$2.55-2.60 per 1,000 images, everything included. Our test run of 250 UK images used $0.012 of platform usage in total (3 Google pages).

Does it need a Google account, an API key or a browser?

No. It reads Google's normal results pages over HTTP through Apify's GOOGLE_SERP proxy (Google answers ordinary datacenter and residential requests with a JavaScript wall). No login, no Google API key, no headless browser. The proxy is chosen automatically; you do not need to configure anything.

Can I search Google Images in other countries and languages?

Yes. Country selects the Google domain and pins the location (Google's uule location parameter), so you get what people in that country see, e.g. more .co.uk sources for a UK search. Language sets Google's interface language (hl). Answers Google localises to the wrong country are detected and retried.

Monitor changes on a schedule

Turn on Only new or changed items and save the input as a task with a Schedule (e.g. daily). The actor remembers which images earlier runs of the same search returned (by Google's image id) and outputs only new images; images seen before are not saved and not charged. Each output image gets changeType: "new" (images do not "change", so previousValues never appears here). Combine it with Time: past week to follow fresh images for a brand, product or topic. Use Monitoring key to keep several schedules on the same search separate; Max images is then the number of images checked per run.

Is it legal to scrape Google Images?

The actor collects publicly visible search results (image links, titles, source pages and the credit lines publishers attach to their images) that Google shows to anyone without logging in; it does not download images. Images found this way are usually copyrighted by their owners: check the licenseUrl / source page before you reuse an image, and check Google's terms and the laws that apply to your use case.

Limits

• Google's Type: Photo and Face filters are gone from its Tools menu, and Google no longer honours itp:photo (tested: the page shows "Any type"); the actor offers the types Google still has.
• Licence fields are filled only for images Google marks as licensable (stock sites, Wikimedia, some photographers); for most images they are null. fileSize is Google's rounded label.
• Depth per search depends on Google (165-300+ images per search in tests); sub-searches go further but the later ones are narrower.
• SafeSearch "Google default" blurs explicit images in the browser but still lists them; choose "Filter explicit results" to remove them.