newgoogle serp api · structured results from one callsee the changelog →

scrape anything.
even the hard stuff

one endpoint. proxies, fingerprints and browser rendering handled. you get clean json, not 400kb of html.

get an api key →./read-the-docs.sh
~/projects/scraper — zsh — 88×24
$ curl api.scantir.com/v1/scrape \ -H "authorization: bearer $KEY" \ -d '{"url":"https://target.com","auto_escalate":true,"extract":{"title":"h1","price":".price"}}' → http tier · residential exit ...... blocked (waf: cloudflare) → escalating to browser tier ....... ok → rendering dom .................... ok ✓ response 200 · method: browser · {"title":"honeycomb jar, 8oz","price":"$42"}
free diagnostics

test the target
before the build.

run public checks for scrapability, WAF signals, structured data, robots.txt, and sitemap discovery.

diagnostics

Scrapability test

Diagnose whether a URL is reachable with direct HTTP, needs browser rendering, or shows anti-bot friction.

access signals

WAF detector

Check headers, cookies, status codes, and page signals for common WAF and CDN fingerprints.

extraction

Structured data extractor

Preview JSON-LD, OpenGraph, canonical tags, and product-like fields before writing an extractor.

compliance

Robots.txt analyzer

Parse robots.txt directives, sitemap links, crawl-delay hints, and user-agent groups for a domain.

discovery

Sitemap extractor

Find sitemap URLs, nested indexes, lastmod values, and sample URLs for catalog discovery.

01 — the toolkit

the infra you'd ship,
if you had six more months.

everything we learned keeping scrapers alive against modern anti-bot, shipped as a single api.

01 / 06

waf detection + escalation

we recognise cloudflare, datadome, akamai and friends from the response. set auto_escalate and a blocked request climbs from plain http to a real browser.

02 / 06

per-site learning

every outcome feeds a per-site profile. the fingerprint and route that worked last time are tried first next time.

03 / 06

css extraction

pass a map of field → css selector. get parsed fields back. skip the html-soup parsing entirely.

04 / 06

residential proxies

managed residential pool, rotated per request. pin the exit country with one field.

05 / 06

js rendering

full chromium. lazy-loaded spas, infinite scroll, shadow dom. wait on selectors, not arbitrary sleeps. screenshots on request.

06 / 06

google serp api

one call returns structured google results with country and language targeting. query / gl / hl aliases make migrating from other serp apis a find-and-replace.

02 — the api

four curl lines
and you're in.

no sdk install. no auth dance. no websockets. just http + json.

  • one endpoint: POST /v1/scrape
  • bearer auth. rotate keys from the console.
  • plain http first; with auto_escalate, a real browser only when the site blocks it
  • curl, python and node samples in the docs
# scrape a js-heavy page, extract the price
curl https://api.scantir.com/v1/scrape \
  -H "authorization: bearer $KEY" \
  -d '{"url":"https://example.com/p/42","render":true,"country":"us","extract":{"price":".price"}}'

# → json with status, method, waf and your extracted data
ready when you are

ship the scraper
on tuesday.

1,000 requests free. no card. no call.

$ get-api-key →./docs
scantir.com · 2026