Serpix — SERP API
Google results in.
Clean JSON out.
Serpix turns live searches into structured, geo-targeted JSON — parsed, typed, and ready for your pipeline. One GET request, no proxies, no headless browsers.
start free · no credit card
Fig. 000 — sample response · live playground ships with the beta
01 — from page to payload
The results page, as a data structure.
Everything you can see on a search results page arrives as typed, documented fields — positions, titles, snippets, ratings, sitelinks, answer boxes. No HTML. No parsing. No selectors to babysit.
We logged 400 miles in this year's releases to find the best trainers for every kind of run…3
Independent lab measurements of stack height, drop and…
Fig. 001 — parse map · one result, fully typed
02 — quickstart
Three lines to your first SERP.
Grab a key, call the endpoint, read organic. That's the whole integration — the response schema is documented, versioned, and stable.
curl "https://api.serpix.com/v1/search?q=coffee+grinders&gl=de" \
-H "Authorization: Bearer $SERPIX_KEY"03 — coverage index
Every surface of the results page.
One key, one schema convention, a growing set of endpoints. Web, image and news search land in v1 — the rest is sequenced on the public roadmap.
Fig. 002 — result surfaces, exploded
Web Search
GET /v1/search
Organic results, snippets, sitelinks, answer boxes, related searches.
v1Image Search
GET /v1/images
Thumbnails, source pages, dimensions and licensing hints per image.
v1News
GET /v1/news
Headlines with sources, timestamps and thumbnails, sorted as ranked.
v1Shopping
GET /v1/shopping
Product listings with prices, merchants, ratings and availability.
roadmapMaps & Local
GET /v1/maps
Local pack results: names, ratings, hours, coordinates, place IDs.
roadmapAutocomplete
GET /v1/autocomplete
Query suggestions as users would see them, per locale and language.
roadmap04 — pricing model
Credits, not subscriptions.
Search volume is spiky. Your bill shouldn't pretend otherwise — meter the requests, skip the retainer.
Fig. 003 — the meter · failed requests spend nothing
Top up when you need
Buy a pack of credits and go. No monthly minimum, no seats, no plan to remember to cancel.
Pay per successful request
A response you can actually use is the only thing that spends credits. Errors and empty pages don't.
Scale without cliffs
Credits buy the same API at every volume — from your first request to your ten-millionth.
full price list ships with the public launch
05 — faq
Questions, answered.
What is Serpix?
Serpix is a SERP API: you send a search query over HTTPS and get back the results page as structured JSON — organic results, snippets, ratings, answer boxes and metadata — without running scrapers, proxies or headless browsers yourself.
How does credit-based pricing work?
You buy credits up front and spend them per successful request — there's no subscription, no monthly minimum and no seat pricing. The detailed credit schedule per endpoint is published at launch.
Do I need to manage proxies or CAPTCHAs?
No. Rotating infrastructure, retries and anti-bot handling are Serpix's job. You call one HTTPS endpoint; if a request can't be fulfilled, it doesn't cost credits.
Which search surfaces are supported?
Web, image and news search ship in v1. Shopping, maps/local and autocomplete are on the public roadmap, all under the same key and schema conventions.
Can I geo-target results?
Yes — pass a country (gl) and language (hl), and results are localized as a user in that market would see them.