How to scrape Google Images with Python
One endpoint, every image as clean JSON, no browser automation.
How do you scrape Google Images?
Scraping Google Images yourself means driving a browser and parsing a grid that loads through JavaScript. Here you send one HTTP request and get clean JSON back.
- 1
Create a free account. Try it free, no card needed.
- 2
Copy your key from the dashboard. One key works for every endpoint.
- 3
Run the curl below, or the Python after it. Every image lands as clean JSON.
First a quick test with curl:
Then the same call in Python with the requests library. Install it with pip install requests, drop in a key, and run:
The results live under the images key. Here are the first two that came back for sunset, shown as CSV so they drop straight into Sheets or a database:
This one query returned 93 images in a single response. To pull more, add a page number and loop, collecting each page into one list:
Change q to any search you would type into the Google Images box, like golden retriever puppy or mid century sofa. Add gl and hl to set the country and language.
What data do you get?
Every image result comes complete, so you build a mood board, a dataset, or a monitoring feed with no second call. Here is one record as CSV, ready for Sheets or a database:
Or the same record as JSON, exactly as the API returns it:
The imageUrl is the full-size file on the source site, thumbnailUrl is Google's preview, and link is the page it sits on. Pixel dimensions come with both, and some results add creator and copyright.
vs writing your own scraper
You can build an image scraper yourself, and plenty of tutorials show how. The trade is maintenance; here is the difference in practice.
| Writing it yourself | CrustAPI | |
|---|---|---|
| Setup | Install a headless browser, handle blocks and page changes | One GET request, no browser |
| Output | Parse the results grid yourself; fields break when the layout changes | Clean JSON from the API, or the same rows as CSV |
| Maintenance | Fix the scraper each time the page changes | We keep it working |
| Free tier | Free, but you pay in time and blocked requests | $6 of free usage each month |
| Cost model | Server, proxy, and developer time | One charge per search; empty searches free |
| Storing the data | Yours to manage | No storage limits. Export to CSV or a database |
How much does scraping images cost?
A search bills the applicable endpoint price no matter how many images come back, and a search that returns none costs nothing.
| Deposit | Google Search / 1,000 requests | Maps / 1,000 businesses | LinkedIn reads / 1,000 requests | People, Jobs, Refresh / 1,000 units |
|---|---|---|---|---|
| $10+ | $1.00 | $1.96 | $6.00 | $1.96 |
| $149+ | $0.76 | $1.49 | $4.50 | $1.49 |
| $549+ | $0.56 | $1.10 | $3.30 | $1.10 |
| $1,999+ | $0.41 | $0.80 | $2.45 | $0.80 |
| $6,500+ | $0.33 | $0.65 | $1.95 | $0.65 |
| $27,500+ | $0.28 | $0.55 | $1.65 | $0.55 |
| $50,000+ | $0.26 | $0.50 | $1.50 | $0.50 |
| $100,000+ | $0.20 | $0.40 | $1.50 | $0.40 |
Every eligible account gets $6 of free usage monthly. Each deposit keeps its prices until spent; paid funds never expire. People and Jobs bill per successful search, Refresh per accessible profile. Full profiles in People use the read rate per full profile. See all billing details.
Larger deposits unlock lower endpoint prices. Each deposit keeps its prices until spent. See the full deposit table on the pricing page. Prices in USD, ex-tax.
Funds are prepaid and never expire, so an unused pack is still yours next year.
Who scrapes Google Images?
- ML and data teams. Build labeled image datasets with dimensions and source links in every row.
- Ecommerce teams. See which product photos rank for a keyword, per country and language.
- Brand watchers. Track where a logo or product shot shows up across the web, query by query.
- Designers and researchers. Pull full-size sources for a mood board in one call.
Save the results to a CSV file
Because the response is plain JSON, Python's built-in csv module writes it to a file in a few lines, no extra libraries. This reads every image and writes one row each:
Open images.csv in Sheets or Excel and you have a finished image index. To download the files themselves, loop over the rows and fetch each imageUrl.
Related: Google Images API · Google Images Scraper · Google Search API · Google News API · API docs
When a different tool fits better: licensed stock imagery, rights-cleared photos, or a contract with a formal SLA are a stock-photo license or a Google Cloud product, not scraping. This endpoint returns the public image results anyone sees on the Google Images page. Check the license on each source page before you reuse a file.
Common questions
No. With the CrustAPI endpoint you send one HTTP request with the requests library and get clean JSON back. There is no browser to drive and no HTML to parse, so the code above is the whole scraper.
Title, source site name, the page link, a direct image URL, a thumbnail URL, and pixel dimensions for both the image and the thumbnail, plus its position in the results. A minority of results also include creator and copyright, when the source publishes them. Enough to build an index or a dataset without a second call.
Logged-out, public-only collection has repeatedly defeated computer-fraud and contract claims. What you get is the same public image results anyone sees. Check each source page's license before reusing a file.
A search returns the image results Google shows for that query. In the sample above one query returned 93 images. You pay one charge per search, and a search that finds none costs nothing.
The response is plain JSON, so Python's built-in csv module writes it to a file in a few lines. The Save to CSV section above has the exact code, no extra libraries.
Yes. Try it free, no credit card, and paid funds never expire. That is enough to pull several full result pages while you build.
Clay
Try it on any search term
the free plan covers a real search. Run one query and look at the JSON that comes back.
Try for free