Google search results, as a plain REST call
Organic results, the AI Overview and the sources it cites — one request, one JSON body. Serper-compatible, so migrating is a change of base URL.
curl --request GET \
--url 'https://serpens.p.rapidapi.com/api/v1/search?q=best%20search%20apis&gl=us&hl=en&format=json' \
--header 'x-rapidapi-host: serpens.p.rapidapi.com' \
--header 'x-rapidapi-key: YOUR_KEY'The numbers above are ours
Everything on this site comes out of the same endpoint we sell. We run 4,185 queries once a week from the United States, and a 300-query sample from nine more countries, store what comes back, and publish the aggregates — 82,620 observations so far, across 12 weeks.
It is the shortest honest answer to the only question that matters when you are buying search data: does it work, and on what. Read how we measure, then check any figure against a run of your own.
What one call gives you
| Field | What it is |
|---|---|
organic[] | Position, title, link, snippet, source and sitelinks for every result on page one |
aiOverview | Google's own answer plus the list of sources it credits — the block most providers do not return at all |
peopleAlsoAsk | The related questions Google renders, each one a query in its own right |
relatedSearches | Google's own map of what sits next to your subject |
knowledgeGraph | Present when Google treats the subject as an entity rather than a phrase |
has_next_page | Pagination without an upper bound. Page 40 is a call like any other |
Optional blocks are absent when Google renders nothing — never null, never an empty array. Test with 'aiOverview' in body and the ambiguity never arises.
Coming from Serper
The response body is the same shape, field for field: organic, knowledgeGraph, peopleAlsoAsk, relatedSearches, sitelinks. No mapping layer, no rewriting your parser. Point your base URL at us and the code you already wrote keeps working.
Three fields are additive — organic[].source, has_next_page, and a dedicated AI Overview endpoint. A client that ignores them behaves exactly as it did before.
Scope, stated plainly: this covers the web search endpoint. Image, news, places, scholar and shopping verticals are not part of the product — if your integration calls those, it needs more than a base-URL change. The full migration note.
Where Google answers for itself
Eight of 24 topics, ranked by how often Google writes the answer above the results. The full index.
Result blocks, by how often they appear
If you are budgeting for a parser, these shares decide how much of it will ever run.
Start on the free tier
A key in under a minute, no card. The free tier is a real allowance, not a trial clock — it does not expire.