Google Images Scraper

Pull image search results from Google Images at scale — full-resolution image URLs, pixel dimensions, file size, source domains, source-page titles, Google-hosted thumbnails, and a docid join key for every tile. Filter by 16 dominant colors, 5 aspect ratios, 6 image types, 6 time windows, 3 license buckets, 3 size brackets, and SafeSearch — across 250 countries and 208 interface languages. Built for creative teams, e-commerce catalogers, and ML dataset builders who need clean, structured image data without clicking through Google one tile at a time.

Why This Scraper?

• 18 typed fields per image — imageUrl, imageWidth, imageHeight, thumbnailUrl, thumbnailWidth, thumbnailHeight, sourceUrl, sourceDomain, sourceTitle, title, imageBytes, docid, position, googleUrl, searchUrl, query, scrapedAt, recordType. No HTML, no post-processing.
• Direct full-resolution image URLs, not just thumbnails — every row carries the original imageUrl with pixel imageWidth x imageHeight and the human-readable imageBytes Google reports (e.g. "214KB").
• 16 dominant-color filters — full color, black & white, transparent background, plus 12 named colors (red, orange, yellow, green, teal, blue, purple, pink, white, gray, black, brown). Match a brand palette in one input.
• Up to 260 unique images per query — hits Google's own per-query ceiling with automatic pagination across start=0/10/20/30 and docid-level deduplication.
• Every Google Images filter wired into a dropdown — 5 aspect ratios (Tall, Square, Wide, Panoramic), 6 image types (Photo, Clip art, Line drawing, GIF, Animated), 6 time windows (hour, 24h, week, month, year), 3 license buckets (Creative Commons, Commercial), and 3 size brackets (Large, Medium, Icon).
• 250 countries and 208 interface languages — full coverage of every regional Google frontend, with the Accept-Language header auto-matched to your language choice so hl= actually changes results.
• includeRelatedQueries for keyword expansion — toggles a bonus recordType: "relatedQueries" row per query carrying Google's "Related searches" chips, ready for trend research.
• Stable docid join key on every row — Google's internal document identifier survives across runs, so you can dedupe, diff, and join scrape outputs over time.
• Paste any Google Images URL — power users can craft a search visually with tbs= filters, copy the URL, and drop it in; every encoded filter is preserved.

Use Cases

Marketing, Creative & Brand

• Build mood boards filtered by dominant color to match a brand palette
• Source on-brand stock candidates by combining imageType: "photo" with a Creative Commons license filter
• Track how a campaign visual is being repurposed across the open web

E-commerce & Product Catalogs

• Pull reference photography for new SKUs from a batch of product names
• Collect competitor packaging shots filtered to imageType: "photo" and aspectRatio: "square"
• Build hero-image candidates for category pages by aspect ratio

Market & Consumer Research

• Run the same query across multiple country codes to map regional visual differences
• Use the includeRelatedQueries output for keyword-expansion research
• Track meme and trend imagery week-over-week with the pastWeek time filter

SEO & Digital Marketing

• Audit which sourceDomain values dominate Google Images for target keywords
• Map the source pages Google rewards with image visibility for branded vs. non-branded terms
• Diff visibility across a 12-month window with the pastYear filter

ML, Computer Vision & Datasets

• Build training sets pre-filtered to Creative Commons with width/height included for resolution gating
• Seed image-recognition pipelines with query-tagged URLs and the docid as a stable join key
• Generate per-class folders by running batched queries with shared filter settings

Getting Started

Simple Search

The fastest way to start — just a keyword:

{
    "queries": ["golden retriever puppies"],
    "maxResultsPerQuery": 100
}

Multi-Query with Filters

Run several queries with the same filters applied to every one:

{
    "queries": ["minimalist kitchen", "scandinavian living room", "japandi bedroom"],
    "maxResultsPerQuery": 200,
    "imageSize": "large",
    "imageType": "photo",
    "aspectRatio": "wide",
    "imageColor": "white",
    "usageRights": "creativeCommons",
    "safeSearch": "moderate",
    "country": "us",
    "language": "en",
    "includeRelatedQueries": true
}

Paste a Google Images URL

Already built the perfect search in Google's UI? Copy the URL from your browser bar and drop it in:

{
    "searchUrls": [
        "https://www.google.com/search?q=sunset+over+mountains&udm=2&tbs=isz:l,itp:photo,ic:specific,isc:orange"
    ],
    "maxResultsPerQuery": 150
}

Any filters already encoded in the URL are preserved exactly as you set them.

Localized Search — Match Country and Language

Pull German-hosted results with a German interface:

{
    "queries": ["kuchen"],
    "country": "de",
    "language": "de",
    "maxResultsPerQuery": 100
}

Input Reference

Search

Parameter	Type	Default	Description
queries	string[]	["golden retriever puppies"]	Keywords to search on Google Images. Each query runs independently and produces its own image set. Up to 100 queries and URLs combined per run.
searchUrls	string[]	[]	Paste full Google Images result URLs. Filters encoded in the URL are preserved verbatim.
maxResultsPerQuery	integer	100	Maximum images per query or URL. Set 0 for all available. Google caps each query at roughly 260 unique images.

Filters

All filters are optional — leave on the defaults for no filter.

Parameter	Type	Default	Options
imageSize	string	any	Any size, Large, Medium, Icon
imageColor	string	any	Any color, Full color, Black and white, Transparent background, Red, Orange, Yellow, Green, Teal, Blue, Purple, Pink, White, Gray, Black, Brown
imageType	string	any	Any type, Photo, Clip art, Line drawing, GIF, Animated
aspectRatio	string	any	Any aspect, Tall (portrait), Square, Wide (landscape), Panoramic
timePeriod	string	anytime	Anytime, Past hour, Past 24 hours, Past week, Past month, Past year
usageRights	string	any	Any license, Creative Commons licenses, Commercial & other licenses
safeSearch	string	off	Off, Moderate, Strict

The usageRights filter surfaces images Google has classified under each license bucket. Always verify the license at the source page before using any image commercially.

Localization

Parameter	Type	Default	Description
language	string	en	Interface language. 208 options spanning English, Spanish (8 regional variants), Portuguese (Brazil, Portugal), Chinese (Simplified, Traditional, Hong Kong), Japanese, Korean, Arabic, Hindi, and 190+ more.
country	string	us	Country code for the regional Google frontend. 250 options covering every country Google supports.

Enrichment

Parameter	Type	Default	Description
includeRelatedQueries	boolean	false	Emit a recordType: "relatedQueries" row per query carrying Google's "Related searches" suggestions.

Output

Every row carries a recordType field — image or relatedQueries — so downstream filtering is one column away.

Image record (recordType: "image")

{
    "recordType": "image",
    "query": "golden retriever puppies",
    "position": 1,
    "title": "Golden Retriever Puppy Playing in the Grass",
    "imageUrl": "https://example-media.com/uploads/puppy-hero.jpg",
    "imageWidth": 1600,
    "imageHeight": 1067,
    "thumbnailUrl": "https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9Gc...",
    "thumbnailWidth": 276,
    "thumbnailHeight": 183,
    "sourceUrl": "https://example-media.com/blog/golden-retriever-puppies",
    "sourceDomain": "example-media.com",
    "sourceTitle": "Golden Retriever Puppies — A Complete Guide",
    "imageBytes": "214KB",
    "googleUrl": "https://www.google.com/imgres?docid=abc123...",
    "searchUrl": "https://www.google.com/search?q=golden+retriever+puppies&udm=2&gl=us&hl=en&start=0",
    "docid": "0wJpF3VkXlABcM",
    "scrapedAt": "2026-04-22T14:32:11Z"
}

Field	Type	Description
recordType	string	Always "image" for image rows
query	string	The keyword that produced this result (or the auto-extracted query for pasted URLs)
position	integer	1-based rank of the image within its query
title	string	Image title or alt text as surfaced by Google
imageUrl	string	Direct URL to the full-resolution image
imageWidth	integer	Full-resolution width in pixels
imageHeight	integer	Full-resolution height in pixels
thumbnailUrl	string	Google-hosted, CDN-stable thumbnail URL
thumbnailWidth	integer	Thumbnail width in pixels
thumbnailHeight	integer	Thumbnail height in pixels
sourceUrl	string	URL of the page where the image was found
sourceDomain	string	Hostname of the source page (e.g. example-media.com)
sourceTitle	string	Title of the source page
imageBytes	string	Human-readable file size as reported by Google (e.g. "214KB")
googleUrl	string	Google's in-product image preview URL
searchUrl	string	The exact Google Images URL this result was scraped from (query + filters + pagination)
docid	string	Google's internal document identifier — stable join key across runs
scrapedAt	string	ISO 8601 UTC timestamp of when the row was captured

Related-queries record (recordType: "relatedQueries")

Emitted only when includeRelatedQueries: true. One row per query.

{
    "recordType": "relatedQueries",
    "query": "nike shoes",
    "relatedQueries": ["Air max", "Air force", "High top", "Sneaker", "Basketball"],
    "scrapedAt": "2026-04-22T14:32:47Z"
}

Field	Type	Description
recordType	string	Always "relatedQueries"
query	string	The source query Google suggested these refinements for
relatedQueries	string[]	Array of Google's "Related searches" suggestions
scrapedAt	string	ISO 8601 UTC timestamp of when the row was captured

Tips for Best Results

• Google caps every query at ~260 unique images — this is a platform-wide ceiling, identical whether you scroll Images yourself or use any tool. To exceed it, split broad queries into specific variants ("nike air max" + "nike air force" + "nike dunk" instead of "nike shoes") and run them as a batch.
• Match country and language — pairing country: "de" with language: "de" returns German-hosted results with a German interface; mismatching the two can silently dilute the result set with cross-locale tiles.
• Use searchUrls for power-user filter combos — Google's UI exposes some filter mixes (e.g. imageColor: "orange" + aspectRatio: "panoramic" + timePeriod: "pastWeek") that are easier to compose visually. Build the search in your browser, copy the URL, paste it in.
• Enable includeRelatedQueries on a topic-discovery run — one extra row per query gives you Google's keyword-expansion list for free, so you don't need a separate run to find adjacent search terms.
• Verify licenses at the source page — the usageRights filter reflects how Google classifies each image, but licensing on the host page can change. Click through to sourceUrl before any commercial use.
• Use docid to dedupe across runs — the same image can re-enter the result set under different queries or on different days. The docid is Google's stable identifier — group on it before merging datasets.
• Start small, then scale — try maxResultsPerQuery: 50 first to confirm the filter combination matches what you want, then bump to 200–260 per query for the production run.

Pricing

$2.00 per 1,000 results — the lowest price on Apify for Google Images data.

Results	Estimated Cost
100	$0.20
1,000	$2.00
10,000	$20.00
100,000	$200.00

A "result" is any row in the output dataset — an image or a related-queries record. Platform fees (compute, storage) are additional and depend on your Apify plan.

Integrations

Export data in JSON, CSV, Excel, XML, or RSS. Connect to 1,500+ apps via:

• Zapier / Make / n8n — Workflow automation
• Google Sheets — Direct spreadsheet export
• Slack / Email — Notifications on new results
• Webhooks — Trigger custom APIs on run completion
• Apify API — Full programmatic access

Legal & Ethical Use

This actor is designed for legitimate market research, competitive analysis, creative inspiration, and dataset building. Users are responsible for complying with applicable laws and Google's Terms of Service. Image copyright belongs to the original creator — always verify licensing at the source page before reusing any image commercially. Do not use extracted data for spam, harassment, copyright infringement, or any illegal purpose.