Google SERP Scraper: AI Overview, PAA and Rank Tracking

Scrape complete Google search results pages as clean JSON, organic results, ads, AI Overview with every cited source, People also ask, local pack, featured snippet, knowledge graph and related searches, plus Google AI Mode, Google News, Google Images and built in rank tracking for your domains. No proxies, no Google account, no setup.

Why this Actor

• Speed: standard processing (the default, no add on) usually delivers results in 1 to 5 minutes; choose Instant under Processing speed for results in seconds at +$0.0018 per page.
• $1.50 per 1,000 result pages ($1.20 on Gold and above) with every SERP feature in the same row, no extra fee for the AI Overview shown on the page, People also ask, local pack, rank tracking or flat result rows.
• Google AI Mode answers for $2.50 per 1,000 ($1.80 on Gold and above): the full answer text, Markdown and all cited sources, with your domain's citation position.
• 32 fields per results page, 22 fields per organic result (position, absolute position, URL, displayed URL, sitelinks, rating, date …).
• 249 countries, 91 languages, city level or GPS level location, desktop or mobile, any Google domain.
• Built to finish runs: busy or network errors are retried automatically (up to 4 attempts with increasing waits); searches rejected before processing are free, searches Google answers with nothing cost only a small no results fee, and every problem is listed with the reason in RUN_SUMMARY; your maximum charge per run is checked before anything is requested, and the run stops cleanly at the limit.

One example row from a real run, shortened:

{
  "id": "482afcf38b538266347d",
  "rowType": "serpPage",
  "searchTerm": "pizza",
  "searchType": "organic",
  "country": "GB",
  "language": "en",
  "location": "London,England,United Kingdom",
  "device": "desktop",
  "googleDomain": "google.com",
  "searchUrl": "https://www.google.com/search?q=pizza&hl=en&gl=GB&ie=UTF-8&uule=w+CAIQIFISCXXeIa8LoNhHEZkq1d1aOpZS",
  "totalResultsEstimate": 1210000000,
  "serpFeatures": [
    "find_results_on",
    "local_pack",
    "organic",
    "people_also_ask",
    "related_searches",
    "google_reviews",
    "knowledge_graph"
  ],
  "scrapedAt": "2026-09-28T06:19:15.000Z",
  "page": 1,
  "organicResults": [
    {
      "position": 4,
      "absolutePosition": 9,
      "title": "30 best pizza restaurants in London 2026",
      "url": "https://www.bbcgoodfood.com/travel/the-best-pizza-in-london",
      "domain": "www.bbcgoodfood.com",
      "displayedUrl": "https://www.bbcgoodfood.com › Travel",
      "websiteName": "Good Food",
      "description": "12 Apr 2026 — Best for a bargain · 1. Vincenzo's · 2. Ria's · 3. Dough Hands · 4. Voodoo Ray's ·  Read more",
      "preSnippet": "04/12/2026 00:00:00",
      "extendedSnippet": null,
      "highlighted": [
        "1. Vincenzo's"
      ],
  …
}

Every dataset row has a stable id and a rowType: serpPage (one Google results page), aiModeAnswer (Google AI Mode), or, when Also output one row per result is on, organicResult, paidResult, newsResult, imageResult or citation.

Results page (rowType: "serpPage")

AI Mode answer (rowType: "aiModeAnswer"): the context fields above plus answerText, answerMarkdown, citations, otherElements and rankTracking (whether and where your domain is cited).

News / Images rows: newsResults (title, url, domain, source, snippet, published, date, imageUrl) or imageResults (title, alt, source, pageUrl, imageUrl, thumbnailUrl, domain), plus resultsCount, pagesCovered and rankTracking.

Flat result rows (free, optional): id, rowType, searchTerm, page, position, absolutePosition, title, url, domain, displayedUrl, description, date, isTrackedDomain, country, language, scrapedAt.

You pay per event, through your Apify account. Main price: $1.50 per 1,000 result pages on the Apify Free plan and $1.20 on Gold, plus $0.002 per run ($0.0006 on Gold). Every event and plan:

Every run is one HTTP request to the Apify API with your Apify token; the input below is the actor's own example. The synchronous endpoint returns the rows in the response for runs up to 300 seconds; for longer runs start a run and fetch its dataset.

curl -X POST "https://api.apify.com/v2/acts/cheapapi~google-serp-scraper/run-sync-get-dataset-items" \
  -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
  -d '{"searchTerms":["best running shoes","how to brew cold brew coffee"],"country":"US","language":"en","maxPagesPerQuery":1,"searchType":"organic","device":"desktop","processingSpeed":"auto"}'

With the JavaScript client:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('cheapapi/google-serp-scraper').call({
  "searchTerms": [
    "best running shoes",
    "how to brew cold brew coffee"
  ],
  "country": "US",
  "language": "en",
  "maxPagesPerQuery": 1,
  "searchType": "organic",
  "device": "desktop",
  "processingSpeed": "auto"
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();

Typical run: 1,000 keywords × 1 results page (about 10,000 organic results). Prices are the Actor fees per listing (Free plan / Gold plan).

Prices from public Apify Store listings, checked September 2026 (start fees of $0.0003 to $0.002 per run left out). ✗ = not mentioned in the listing. Apify platform usage is billed separately by Apify for this Actor too (see Pricing); the table compares Actor fees only.

Per plan, 1,000 result pages (most used Actor vs. this Actor). Apify lets us set Free, Bronze, Silver and Gold prices; Platinum and Diamond pay our Gold price.

Plainly: on Silver we are 23% cheaper, not more; on Gold and above our page price matches the most used Actor's Diamond price, so a Diamond customer pays the same per page (their AI Overview add on, $0.50 per 1,000 on Diamond, is included here).

Honest take: if you only need position, title, URL and snippet of organic results, the cheapest per page Actor costs about a third of our price, and a per result Actor is similar on Gold. You pay $1.50 here for the full page in one row, AI Overview with cited sources, People also ask, local pack, ads, featured snippet, knowledge graph, plus rank tracking of your domains and optional Google AI Mode, at no add on fee for the AI Overview Google shows on the page.

• Rank tracking: schedule daily or weekly runs for your keywords and domains and follow position, page and the ranking URL over time.
• AI search visibility (GEO): see whether your domain is cited in Google's AI Overview and AI Mode answers, and at which citation position.
• Content research: collect People also ask questions, related searches and featured snippets to plan articles and FAQs.
• Local SEO: check the local pack for a city or GPS point, with ratings, review counts and phone numbers of the businesses shown.
• Competitor and ads monitoring: see who ranks and who advertises (paidResults) on your commercial keywords, per country and device.
• LLM and RAG pipelines: feed fresh Google results, AI answers and cited sources to your AI agents.

• AI answers from other engines (ChatGPT, Perplexity, Gemini, Claude), this Actor covers Google only (AI Overview and AI Mode). For cross engine AI answers and brand citations use our AI Search Visibility Tracker.
• Full Google Maps place data (opening hours, all reviews, photos) beyond the local pack, use our Google Maps Scraper Plus.
• Search volume, keyword difficulty and keyword ideas, use our Keyword Research Tool.
• Content of the ranking pages, contact/lead enrichment, Google Shopping product pages and screenshots of the results page, not collected; the Actor returns what Google shows on the results page.

• Results reflect Google at the moment of collection. Google personalizes and tests results constantly, so small differences from your own browser are normal.
• Google shows about 10 organic results per page, sometimes fewer. Depth is limited to 700 results per search term (Google Search, News and Images).
• AI Mode returns one answer per search term (no result pages) and cannot use Google search URLs. News is always collected as desktop.
• Embedded images inside AI answers are not included (the text, links and all cited sources are).
• With Standard speed, runs at busy times can take longer than 5 minutes; searches not finished within "Maximum wait" are reported in RUN_SUMMARY and charged the no results fee. If the platform moves a run to another server, searches already sent are collected, never sent twice.
• Some fields (ratings, dates) are null when Google does not show them. People also ask questions are always collected, but Google now loads most answers as AI generated text only after a click, so answer is often null (answerIsAiGenerated: true) even with expansion on, in our September 2026 test runs every answer was null. Expansion reveals extra questions, not answers.