Google Images Scraper - Image Search Results & Direct URLs

Scrape Google Images at scale - direct full-size image URLs, dimensions and the page each image sits on. Up to 100 images per request, so image datasets build quickly and cheaply.

⭐ up to 100 images per page · direct full-size URLs · country + language + date filters · pay per row delivered · CSV/JSON/Excel · API, MCP & schedules

What data does it extract?

One row per image, with the query, page and type it came from so multi-query exports stay separable. Fields Google does not return for an item are simply absent, and any new field Google adds is passed through automatically.

Field	Description
title	Image title or alt text
imageUrl	Direct full-size image URL
imageWidth	Width in pixels
imageHeight	Height in pixels
thumbnailUrl	Google thumbnail URL
thumbnailWidth	Thumbnail width
thumbnailHeight	Thumbnail height
source	Name of the site hosting the image
domain	Host domain
link	Page the image appears on
position	Rank within the page
query	The keyword this row came from
page	Which results page it came from

Example output

{
  "query": "coffee shops new york",
  "page": 1,
  "type": "image",
  "title": "Original Cafés in NYC and Where to Find Them - VIE Magazine",
  "imageUrl": "https://viemagazine.com/wp-content/uploads/2018/03/vie-magazine-cafes-of-nyc-hero-min.jpg",
  "imageWidth": 1900,
  "imageHeight": 1200,
  "thumbnailUrl": "https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSGYwHGFCH9J2ScD-dVU0_P2FCHvxskBx2ztApFZWKsTw&s=10",
  "thumbnailWidth": 697,
  "thumbnailHeight": 439,
  "source": "VIE Magazine",
  "domain": "viemagazine.com"
}

How to use it

1. Paste your queries into queries - one per line.
2. Set the options you need from the table below.
3. Run it, then download the dataset as CSV, Excel, JSON, JSONL or XML - or read it straight from the Apify API.

Input example

{
  "queries": [
    "latte art",
    "espresso machine"
  ],
  "num": 100,
  "gl": "us",
  "hl": "en",
  "maxPages": 2
}

Option	Type	Description
queries	array	Required. Keywords, one per line.
num	integer	Images per page, up to 100. Default 100.
gl	string	Country code - us, in, gb, de...
hl	string	Language code - en, hi, es...
dateRange	string	qdr:h, qdr:d, qdr:w, qdr:m, qdr:y - past hour to past year.
maxPages	integer	Pages per keyword, default 1. Each page is one request.

Pricing

Event	Price
image-result	$0.99 per 1,000 images ($0.00099 per image)

1,000 images → $0.99 · 10,000 → $9.90 · 100,000 → $99.00

You are billed per row delivered, as it is delivered. Queries that fail or return nothing cost you nothing, and an interrupted run only ever bills the rows already in your dataset.

Use cases

Build image datasets for machine learning or moderation

A single run returns hundreds of direct image URLs with dimensions, ready to download and label. imageWidth/imageHeight let you filter out thumbnails before fetching.

Monitor where your brand, logo or product photos appear

Search your brand terms and watch the domain column for sites republishing your imagery.

Research competitor visual positioning

See which images rank for your category keywords and which domains own that visual real estate.

Source reference imagery for design and content teams

Collect candidate visuals with their source page so attribution and licensing checks stay possible.

Audit image SEO for a topic

Find which pages win image search for your keywords, then compare their alt text and file sizes.

Good to know

• num genuinely works here - up to 100 images per request, unlike Google's text search.
• dateRange changes which images Google returns.
• Image URLs point at third-party sites; check licensing before reuse.

Frequently asked questions

Do I need proxies, an API key or a Google account?

No. Requests run through hosted Google search infrastructure on our side - nothing to configure.

How many images can I get per keyword?

Up to 100 per page via num, and maxPages goes deeper - so 100 keywords at 2 pages yields roughly 20,000 images. Each page is one request.

Do I get the full-size image or just a thumbnail?

Both. imageUrl is the direct full-size file on the source site, and thumbnailUrl is Google's cached preview.

Can I filter images by recency?

Yes - dateRange accepts qdr:h through qdr:y and genuinely changes which images Google returns.

Does it download the image files?

No. It returns URLs and metadata, which keeps runs fast and cheap. Download the files yourself from imageUrl, or chain a download step after the run.

How do I export Google Images results to CSV or Excel?

Download the run's dataset as CSV, Excel, JSON, JSONL or XML, or read it via the Apify API.

Am I charged for keywords that return no images?

No. Billing is per row delivered.

Related Google scrapers

• Google Search Results Scraper - organic rankings and People Also Ask
• Google Videos Scraper - video results across YouTube, TikTok and the web
• Google Places Scraper - local business results with phone and website
• Google Maps Scraper - map places with GPS, hours and Place ID
• Google Patents Scraper - patents with inventors, assignees and dates
• Google Shopping Scraper - products with prices, sellers and ratings

---

Not affiliated with or endorsed by Google. Returns publicly available search results.