SERP API
What SERP API means
A SERP API takes a query and settings such as country, language, and page, loads the search engine's results page, and returns it as JSON: organic results with their positions, plus blocks such as the knowledge graph, People also ask, and local results. It exists because search engines do not offer their own results as an API, and scraping them yourself means handling blocks, CAPTCHAs, and a page layout that changes.
SERP APIs differ in what they parse, how fast they answer, whether they cache, and how they price. Some return cached pages for repeated queries; others fetch every call live. For rank tracking and research, the settings a result was collected with matter as much as the result.
Clair's Google Search API is a SERP API for google.com that fetches a live page per call and echoes the query settings in each response.
How a SERP API works
- You send a query with its settings: the search terms, the country and language to search from, and the page of results.
- The provider loads the results page from the search engine through its own network of addresses, often in a real browser, and retries by another route when the engine blocks the request.
- It parses the page into fields: each organic result with its position, title, URL, and snippet, and each block the engine showed, such as ads, the local pack, or People also ask.
- You get JSON back, usually in a few seconds, and pay per request. Most providers do not charge for failed requests.
What to compare between SERP APIs
- Which blocks it parses. Every API returns organic results; coverage of the knowledge panel, local pack, videos, and AI Overviews varies.
- Speed, at the median and at the slow end. A batch job tolerates a slow 95th percentile; a chatbot waiting on a search does not.
- Success rate, and whether failed requests are billed.
- Location control: country and language only, or also city and coordinates. Local rankings need the second.
- Live or cached. Some APIs return a stored copy when the same query was run recently, which is cheaper and faster but not today's page.
- Price per 1,000 searches at your volume, and whether unused searches carry over.
Three SERP APIs on the same queries
Speed and reliability are the numbers that differ most between SERP APIs, and they are rarely published side by side. We ran the same queries through Clair and two other SERP APIs, one at a time, and kept every response. Clair was the slowest of the three and the only one with failures in this run; the table shows it as measured.
| Succeeded | Median time | 95th percentile | Organic results | |
|---|---|---|---|---|
| Clair | 18 of 20 | 6.2 s | 25.2 s | 7.6 |
| SerpApi | 20 of 20 | 1.5 s | 3.7 s | 7.5 |
| Serper | 20 of 20 | 1.1 s | 1.4 s | 9.1 |
SERP APIs for rank tracking
A rank tracker is the most common use of a SERP API: the same keywords, searched on a schedule, with the position of each domain recorded. For the positions to be comparable from one day to the next, every search needs the same country, language, and device, and every reading needs its date.
Depth costs requests. Google stopped honouring the num=100 parameter in September 2025, so each page of ten results is one request: tracking the top 100 for a keyword takes ten requests, the top 10 takes one. Check that the API counts position across pages, so the first result on page 2 is position 11, not 1.
The Google rank tracking use case on this site builds a tracker on Clair's Google Search API, with the code and the monthly cost at several keyword counts.
How it appears in a response
GET /v1/google-search?q=albert+einstein&country=us&page=1 · results.0
{
"position": 1,
"title": "Albert Einstein",
"url": "https://en.wikipedia.org/wiki/Albert_Einstein",
"source": "Wikipedia",
"displayed_url": "https://en.wikipedia.org › wiki › Albert_Einstein",
"date": null,
"snippet": "Albert Einstein (14 March 1879 – 18 April 1955) was a German-born theoretical physicist ...",
"duration": null,
"sitelinks": []
}Common questions
Why not scrape Google yourself?
You can, at small scale. At volume Google blocks repeated requests from the same addresses and shows CAPTCHAs, and its page markup changes often, so a scraper needs proxies, browser handling, and regular parser fixes. A SERP API does that work and charges per search.
Does Google have an official API for its search results?
Not for the google.com results page. Google's Custom Search JSON API searches a configured engine rather than google.com itself, is closed to new customers, and shuts down on January 1, 2027.
How much does a SERP API cost?
Most charge per search in monthly plans, and prices per 1,000 searches fall as volume rises. The Clair vs SerpApi comparison prices both from 1,000 to 100,000 searches a month.

Glassdoor API
Crunchbase API
Instagram API
Bloomberg API
Associated Press API