SERPs with Pixel Position and AI Overviews

SERP API

The most reliable API for retrieving Google search results.
Fresh data and pixel positions. AI Overviews included.

  • 1,000 free SERPs
  • No credit card
google.fr/search?q=comparateur+assurance+auto
AI Overview
lelynx.frlesfurets.comlecomparateurassurance.com
228 px
1
www.lelynx.fr
Assurance auto : Comparateur et Devis gratuit en ligne
638 px
2
www.lesfurets.com
Assurance Auto : Comparateur & Devis Gratuit en Ligne
808 px
3
mon-auto-assure.com
Comparateur assurance auto pas chère en ligne
978 px
{ "rank_type": 1, "page": 1, "pixel": 638 }

Real example: “comparateur assurance auto”, Google France, desktop

€0.30 per 1,000 SERPs, page 1
No subscription, pay as you go
10 pages per keyword, with Google's real pagination
Desktop and mobile by country, by language and down to the city
Search and Maps two engines, one API

Beyond rankings, the full richness of the SERP

Organic results, AI Overviews, images, related searches… Semscraper collects every element of Google's result pages and returns them as structured data, with their ranking and pixel position. Here is a real example.

response.json
{
  "type": "organic",
  "items": [
    {
      "rank_type": 1,
      "rank_serp": 17,
      "page": 1,
      "pixel": 638,
      "domain": "www.lelynx.fr",
      "title": "Assurance auto : Comparateur et Devis gratuit en ligne"
    }
  ]
}
Simplified excerpt from the JSON response for the search “comparateur assurance auto” on Google France.
rank_type
Rank among results of the same type.
rank_serp
Rank among every element on the page, in display order, AI Overview included.
page
The Google page the result appears on, taken from Google's real pagination.
pixel
Distance in pixels between the top of the page and the result.
In the same response 97 organic results 11 AI Overview citations 6 images 8 related searches

One request, the full richness of the SERP, ready to use.

Data quality

We are proud of the quality of our data

Google changes its rules regularly. We watch, we fix, and you get a SERP that matches what the searcher sees.

The right location

Every SERP is collected from the requested country and city. When Google stopped trusting the gl parameter in May 2026, our fix was live on May 26.

A real browser

Pages are rendered like a searcher's: asynchronous AI Overviews and late-loading blocks are captured, and pixel position is measured on the rendering.

Live collection

No cache: every request triggers a new collection on Google, and only pages actually found are billed.

Integration

An easy-to-use API

1 Quick and easy integration

No code? Connect Semscraper to Claude, ChatGPT or Cursor with the MCP server. For AI agents, RAG or evaluations: the SERP API for AI.

2 Output formats: JSON or HTML
HTML
JSON
Pricing

Pay per request

No subscription: you top up your balance, and each SERP retrieved is deducted. Same price for Google Search and Google Maps.

€0.30 per 1,000 keywords, page 1
+€0.20 per extra page, for 1,000 keywords

1,000 free SERPs when you sign up. Only pages found are billed, and your balance never expires.

FAQ

Frequently asked questions

What is a SERP API?
A SERP API collects a search engine's results page (the SERP) for a keyword on your behalf, and returns it as structured data in JSON. You get every result with its type, rank and position, without handling proxies, captchas or a parser. Semscraper covers Google Search and Google Maps.
Which countries and languages are supported?
186 Google versions, from google.com to google.com.au, and 117 languages, for 372 country-language combinations. You can also geolocate the search down to the city. The full list is available in the API documentation.
Can I choose the location (country, city) and device (desktop, mobile) for my searches?
Absolutely. You can target a country, a department or a city (and its district). You can also choose the scrape medium (desktop or mobile), so you get results that are as close to reality as possible.
What types of results can I retrieve (organic SEO, Google Ads, Maps, Featured Snippets, etc.)?
You can scrape all the elements present in a Google SERP:
- Organic SEO results
- Presence of an AI Overview
- Local Pack / Google Maps
- Featured Snippets
- People Also Ask (PAA)
- Images, videos, Google Shopping
- Related searches
- And more

This gives you a complete overview of the competition and visibility on Google.
How long does it take to process a keyword?
When a keyword is added, it enters a queue specific to your account. Each account has its own queue, independent of the others. So even if a customer has a million keywords in process, it won't impact your turnaround time.

We have a large number of robots working in parallel: each one starts by processing the first keyword for each account, guaranteeing fair distribution and rapid turnaround times.

Our infrastructure is capable of processing tens of millions of scrapes per day, and we have the capacity to scale in a matter of minutes to absorb peaks in demand and exceed this capacity.

A keyword takes a median of about twenty seconds of collection for one results page, and about fifty for 10 pages (measured on 12 million SERPs, 5.3 million one-page and 260,000 ten-page, from September 12 to October 10, 2026), not counting the wait in your queue. A large batch sent at once flows through gradually in your own queue.
Are the recovered data reliable and complete?
Yes, the results come directly from Google, just as a user would see them in their browser. We use real browser instances, which guarantees the authenticity of the data.

What's more, you can geolocate your searches to simulate the SERP as it appears according to the user's precise location (country, city, even neighborhood). This ensures a faithful, contextualized view of the results, bearing in mind that SERPs can differ depending on the time of day and the Google datacenter used.
What data formats are available?
Results are available via :
- JSON (quick and easy to integrate)
- Full HTML SERP

Via the API, you can retrieve one or both formats at no extra cost.
How much does the service cost and how is it billed (per request, per volume, etc.)?
Billing works in a simple and flexible way: you credit your balance, then each request to retrieve a SERP is deducted from this balance. So you only pay for what you actually use.

Rates are degressive according to the amount credited, so you can benefit from lower prices if you top up your account with a larger sum.

We accept payment by credit card, and you can also set up an automatic top-up to ensure that your scrapes continue without interruption.

Ready to integrate our API?

Start for free and get your first SERPs results in less than 5 minutes