Google Search Scraper — SERP results, rank tracking, no browser

Scrape Google organic search results for any list of keywords and export them as JSON, CSV or Excel. Every row gives you organic rank, title, URL and snippet, with country, language and time-range targeting.

Built for SEO monitoring and rank tracking: schedule it, diff position over time, and watch your keywords move. No browser, no API key, nothing to configure.

---

What you can do with it

• Keyword rank tracking — run daily on your target keywords and chart position over time
• SERP monitoring — see who is ranking, who moved, and who appeared
• Competitor analysis — find every page a competitor ranks for using site: queries
• Local SEO — the same keyword in us, gb, de returns different rankings; compare them
• Content research — pull the top results and snippets for a topic cluster
• Lead sourcing — turn site: and intent queries into domain lists, then feed them to a contact scraper
• Feed a RAG pipeline or LLM with fresh, ranked search results

---

How to scrape Google search results

Paste your search terms into Search queries, one per line, and press Start. That is the whole setup.

Google search operators all work:

Operator	Example
Site restriction	site:apify.com actor
Exact phrase	"web scraping tools"
Exclusion	python scraper -selenium
Either/or	serp api OR serp scraper
Title match	intitle:pricing
File type	filetype:pdf market report

---

Input

Field	What it does
Search queries	One per line. Operators supported
Results per query	How many organic results to return. 10 = one request
Country	2-letter code (us, gb, de) — changes the ranking Google serves
Language	2-letter interface language code
Time range	Past hour, 24 hours, week, month, year, or any time
SafeSearch	Google's SafeSearch filtering
Max concurrency	How many queries to run at once

---

Output

One row per organic result. Export as JSON, CSV, Excel, XML or RSS.

{
  "query": "web scraping tools",
  "position": 3,
  "title": "5 Best Web Scraping Tools (Ranked By Real Tests)",
  "url": "https://www.scrapingdog.com/blog",
  "url_is_exact": false,
  "displayed_url": "https://www.scrapingdog.com › Blog",
  "snippet": "Jan 25, 2026 — Scrapingdog - Best overall for scalable, cost..."
}

Field	Notes
query	Which search this result came from, so multi-keyword runs stay separable
position	True organic rank. Ads, AI overviews, shopping carousels and video shelves are excluded, not filtered afterwards
title	The result headline
url	Absolute URL — see the section below on how it is derived
url_is_exact	Whether the URL is verbatim or reconstructed
displayed_url	Google's own breadcrumb form, verbatim and never guessed
snippet	The description text beneath the result

The same URL ranking for two different queries is two rows, because it is two rankings. Dedup is per query, not global.

---

About the URL field — read this one

Google now encrypts its outbound result links. Every link on a modern results page points at /goto?url=<encrypted blob>; the blob is opaque, the tracking attribute repeats the same blob, and requesting it directly returns an error. The real destination is genuinely no longer in the page.

What is in the page is the URL Google displays to a human — https://github.com › apify. So:

• displayed_url is exactly what Google showed. Never guessed.
• url is that turned back into a usable link.
• url_is_exact tells you which kind you got:
  • true — Google displayed a bare domain, and the URL is verbatim
  • false — the path was rebuilt from breadcrumbs. The domain is always right; the path is usually right but can be off, because Google prettifies breadcrumbs and truncates long ones. When a path is truncated, only the origin is returned rather than a URL containing a literal ...
  • null — the result showed no URL at all, typically a video

Filter on url_is_exact when you need certainty. Measured across 20 results: 5 exact, 12 reconstructed, 3 with no URL. Any scraper claiming exact Google URLs today is either following every link individually — at 10x the requests — or guessing without telling you.

---

Pricing

$0.15 per 1,000 results, plus Apify platform usage.

Plan tier	Price per 1,000 results
Free / Bronze	$0.17
Silver, Gold, Platinum, Diamond	$0.15

Read the usage half before you scale this one up. Google serves no results at all to an ordinary HTTP client, so every request goes through Apify's SERP proxy, billed separately at $2.50 per 1,000 SERPs on Free and Starter. One request returns 10 results, so that works out at about $0.25 per 1,000 results — larger than the scraper charge itself, and the biggest line in any run. It falls to $2.00 per 1,000 SERPs on Scale and $1.70 on Business.

Compute is negligible next to that (~$0.01 per 1,000 results): there is no browser and no per-query API fee.

Requests cost money, not results. 10 results is one request, 100 results is ten. Set Results per query to exactly what you need.

---

FAQ

Why can't I just fetch google.com/search myself? You can, and you will get a page with no results in it. Google retired its no-JavaScript results page. A direct fetch returns a "please enable JavaScript" redirect; an older browser signature gets the cookie consent wall instead. Nine user-agent and parameter combinations were tested and every one returned HTTP 200 with zero results. This Actor routes through Apify's SERP proxy, which is what makes plain HTTP work at all.

Does it use a headless browser? No — and that is deliberate. A headless Chromium version was built and measured first: it timed out at 75 seconds per query and needed a ~1.5 GB image. Plain HTTP through the SERP proxy is 1.3–4 seconds per request and far cheaper.

Do I need to configure a proxy? No. It is fixed to the one group that works. A country set in the proxy options is honoured; the group is not changeable, because nothing else returns results.

Are ads included? No. position is organic rank only.

Can I get more than 100 results per keyword? Yes, but Google's own result depth thins out fast and quality drops. Raise Results per query; each extra page is one more request.

Why is a snippet occasionally empty? Some result types genuinely carry no description text. Coverage was 30/30 in the last measured run.

How fast is it? Measured on the platform: 60 results across 3 queries in 13 seconds.

Can I run it on a schedule? Yes — that is the main use. Run daily on a keyword list and diff position to track rankings.

---

Related scrapers

• Website Contact Scraper — turn the domains you find here into emails, phone numbers and social profiles
• LinkedIn Jobs Scraper — job postings without a login or cookie
• TikTok Ad Library Scraper, Meta Ads Library Scraper, Google Ads Transparency Center Scraper — competitor ad intelligence