Google Search Scraper: organic SERP results for rank tracking at scale

What are the Google Search Scraper specifications?

Google Search Scraper accepts keywords or Google search URLs, targets 240 countries and 36 languages, returns desktop or mobile SERPs, and pages as deep as you set. Each result is a normalised record of title, link, visible link and snippet.

What does Google Search Scraper output look like?

Google Search Scraper output is one JSON object per organic result, carrying four fields: title, link, visible_link and snippet, plus a type discriminator. It is deliberately small. Every field is one an SEO report actually uses.

{
  "searchTerms": ["project management software"],
  "countryCode": "us",
  "languageCode": "en",
  "maxPagesPerQuery": 1
}

{
  "type": "searchResult",
  "title": "Welcome to Python.org",
  "link": "https://www.python.org/",
  "visible_link": "www.python.org",
  "snippet": "A programming language that lets you work quickly."
}

{
  "searchTerms": [
    "project management software",
    "task tracker for teams",
    "gantt chart tool"
  ],
  "countryCode": "gb",
  "maxPagesPerQuery": 1
}

{
  "type": "searchResult",
  "title": "Welcome to Python.org",
  "link": "https://www.python.org/",
  "visible_link": "www.python.org",
  "snippet": ""
}

{
  "startUrls": [
    "https://www.google.com/search?q=site%3Aexample.com+pricing"
  ],
  "countryCode": "us",
  "maxPagesPerQuery": 5
}

{
  "type": "searchResult",
  "title": "Pricing, Example",
  "link": "https://example.com/pricing",
  "visible_link": "example.com › pricing",
  "snippet": "Plans from $0 to $99 a month."
}

(object) => ({
 ...object,
  title: object.title?.toUpperCase() || null,
  sourceDomain: object.visible_link || null,
  snippet: undefined
})

{
  "type": "searchResult",
  "title": "WELCOME TO PYTHON.ORG",
  "link": "https://www.python.org/",
  "visible_link": "www.python.org",
  "sourceDomain": "www.python.org"
}

What the fields are

• The result

  Title as Google renders it, and the destination URL

• The display URL

  visible_link, the breadcrumb Google shows, which is not always the link

• The snippet

  Google’s description text, empty when Google renders none

• What you add

  Position from dataset order, and any field your map function returns

How do you get the ranking position?

Google Search Scraper does not return a position field. Rank is the order of the dataset: the first record for a query is position 1. Number the records as they arrive and you have a rank column, four lines of code, and the reason this page says so is that nothing on the actor’s own listing does.

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

TERM = "project management software"

run = client.actor("apidojo/google-search-scraper").call(run_input={
    "searchTerms": [TERM],
    "countryCode": "us",
    "languageCode": "en",
    "maxPagesPerQuery": 1,
})

# Rank is position in the dataset. Enumerate in iteration order, do not sort.
for position, item in enumerate(
    client.dataset(run.default_dataset_id).iterate_items(), start=1
):
    print(position, item["visible_link"], item["link"])

One caveat worth stating plainly: because ads and the paid or feature blocks are not returned, this position is the organic rank, not where the result sits on the rendered page. For share-of-voice work that distinction matters, and no scraper that returns only organic results can close it.

How much does Google Search Scraper cost?

Google Search Scraper costs $0.002 per query with the first 10 results included, then $0.0002 per further result. Checking 1,000 keywords to depth ten costs $2.00. The formula is $0.002 × queries + $0.0002 × (results − 10 × queries).

The README’s own examples for "1,000 results" and "5 keywords, 500 results" do not match its formula, they subtract 100 free results instead of 10. The figures above are computed from the event prices the Apify Store API reports, which are the ones you are billed at. Free plan runs in demo mode: 5 runs a month, 10 items a run.

How does it compare with Google’s Custom Search JSON API?

Google’s Custom Search JSON API is closed to new customers, and Google’s own documentation gives existing customers until 1 January 2027 to move to something else. While it lasts it allows 100 free queries a day, charges $5 per 1,000 after that, caps at 10,000 queries a day, and never returns more than 100 results for a query.

Every figure in the right-hand column is quoted from Google’s Custom Search JSON API overview and its cse.list reference, read on the snapshot date. The overview carries the closure notice and the 1 January 2027 date; the reference states that num caps at 10 and that "the JSON API will never return more than 100 results, even if more than 100 documents match the query". Google names Vertex AI Search as the successor for searching up to 50 domains, and asks anyone who needs full web search to contact them.

How to scrape Google Search with Apidojo in 4 steps

Scraping Google with Apidojo takes 4 steps: open the scraper, paste keywords or a search URL, set the country and depth, then run and export. No API key, Cloud project or search engine configuration is involved.

1. 01

   Open Google Search Scraper on Apify

   Sign in with a free Apify account. No credit card is needed to run it in demo mode.

2. 02

   Paste keywords or a search URL

   Put terms in searchTerms, or paste Google search URLs into startUrls to keep operators such as site: or a tbs filter.

3. 03

   Set country, language and depth

   countryCode is required and also picks the Google domain. Add languageCode, switch mobileResults on for the mobile SERP, and set maxPagesPerQuery for depth.

4. 04

   Run and export the dataset

   Download CSV for a spreadsheet, or read the results through the Apify API and number them to get rank.

How do you call the Google Search Scraper API?

Google Search Scraper is the actor apidojo/google-search-scraper, called through the Apify REST API or the official Python and JavaScript clients. A rank check is three fields: the term, the country, and one page of depth.

from apify_client import ApifyClient
from datetime import date

client = ApifyClient("YOUR_APIFY_TOKEN")

TERMS = ["project management software", "task tracker for teams"]
DOMAIN = "example.com"

run = client.actor("apidojo/google-search-scraper").call(run_input={
    "searchTerms": TERMS,
    "countryCode": "us",
    "languageCode": "en",
    "maxPagesPerQuery": 1,
})

# 20 records: ten for the first term, then ten for the second.
rows = list(client.dataset(run.default_dataset_id).iterate_items())
for i, term in enumerate(TERMS):
    page = rows[i * 10:(i + 1) * 10]
    hit = next((n for n, r in enumerate(page, 1) if DOMAIN in r["link"]), None)
    print(date.today(), term, hit or "not in top 10")

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });

// Who owns the top ten across a keyword set
const run = await client.actor('apidojo/google-search-scraper').call({
  searchTerms: ['serp api', 'google scraper', 'rank tracking api'],
  countryCode: 'us',
  maxPagesPerQuery: 1,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
const share = {};
for (const r of items) {
  const host = new URL(r.link).hostname.replace(/^www\./, '');
  share[host] = (share[host] ?? 0) + 1;
}
console.log(Object.entries(share).sort((a, b) => b[1] - a[1]).slice(0, 10));

curl -X POST \
  "https://api.apify.com/v2/acts/apidojo~google-search-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"searchTerms":["serp api"],"countryCode":"us","languageCode":"en","maxPagesPerQuery":1}'

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com/?tools=apidojo/google-search-scraper",
      "headers": { "Authorization": "Bearer YOUR_APIFY_TOKEN" }
    }
  }
}

• n8n

  A weekly keyword run writing into a rank table.

• Looker Studio

  CSV export as a ranking dashboard source.

• Warehouse

  JSONL into BigQuery, one row per result per day.

• Webhooks

  POST to any URL when a run finishes.

What can you build with Google SERP data?

SERP data powers 6 SEO workflows: rank tracking, share of voice, competitor content mapping, keyword research, per-market comparison and citation monitoring. Each one names the input, the field it reads and what comes out.

• 01

  Rank tracking

  Your keyword set · one page · daily or weekly

  Fields

  link matched against your domain, position from dataset order

  Lands in

  A rank table with date, country and device on every row

• 02

  Share of voice

  A keyword set · top ten · one country

  Fields

  visible_link, counted by host

  Lands in

  Who owns the category, as a percentage of top-ten slots

• 03

  Competitor content mapping

  startUrls with site:competitor.com

  Fields

  title and link for every indexed page Google will show

  Lands in

  Their content inventory, and the gaps in yours

• 04

  Keyword research at ten a penny

  A long keyword list · one page each

  Fields

  title and snippet, read for intent and phrasing

  Lands in

  Intent clusters, 1,000 keywords for $2.00

• 05

  Per-market comparison

  The same keyword × countryCode us, gb, de, tr

  Fields

  overlap between the top tens

  Lands in

  Which markets need their own page, not a translation

• 06

  Citation monitoring

  Brand and product terms · weekly

  Fields

  link and snippet, watched for the sources that rank above you

  Lands in

  A list of the pages worth pitching, reviewing or reclaiming

What are the limits of Google Search Scraper?

Google Search Scraper returns organic results and nothing else. Six limits below matter before you plan a report around it, the first two are the ones people discover late.

• No SERP features at all

  Ads, shopping, the local pack, People Also Ask, featured snippets, knowledge panels, video carousels and AI Overviews are not in the output. What comes back is the organic list.

  Treat the position as organic rank. For pixel position or feature presence you need a rendered-SERP product, which this is not.

• No position field

  Rank is the order of the dataset, not a value on the record. Anything that reorders the rows before you number them loses it.

  Enumerate on read. The rank section above has the four lines.

• No result count and no date

  Google’s "about 12,300,000 results" line and any result timestamp are not returned. Neither is the query, so a batch run does not label which term a row belongs to.

  Stamp the term, country, device and run date on each row with customMapFunction, or track them by position in the batch.

• snippet can be empty

  Google renders no description for some results, and the field comes back as an empty string rather than being absent.

  Test for empty, not for missing, before you write it into a report.

• Roughly 200 results per query

  That is what Google itself serves for a query, so paging deeper returns nothing new however high maxPagesPerQuery goes.

  For more coverage, split the query, by site:, by country, or into narrower terms.

• Personalisation and volatility

  Two runs minutes apart can differ. Google varies results by data centre, time and test bucket, and the actor holds no session that would stabilise them.

  Measure on a fixed schedule and compare trends, never two single runs.

How reliable is Google Search Scraper?

Google Search Scraper ran 11,744 times in the last 30 days at a 95.8% success rate. That is the lowest success rate in the Apidojo catalogue, which averages 99.7%, and it is published here rather than left off the page.

Success is succeeded ÷ succeeded + failed + timed out; 481 of 11,744 runs failed. Google defends its result pages more actively than any other source in the catalogue, and this is what that costs. Plan retries into anything scheduled, and check the run log rather than assuming an empty dataset means an empty SERP. Read on 21 September 2026.

Is it legal to scrape Google search results?

Collecting publicly visible search results without logging in is generally lawful in the United States. US courts have held that scraping public pages does not breach the Computer Fraud and Abuse Act, in hiQ Labs v. LinkedIn (9th Cir., 2022). Google’s Terms of Service are a separate, contractual matter from that statute.

• What the scraper reads

  Public, logged-out result pages. No Google account, no session and no authentication is involved, so nothing gated is reached.

• Terms of Service are not the CFAA

  hiQ addressed a criminal statute, not a contract. Google’s terms discourage automated querying, and that risk is a commercial judgement for you to make, not one this page can make for you.

• Snippets are other people’s text

  A snippet is an extract of somebody’s page, and titles are theirs too. Analysing them is one thing; republishing a corpus of them is a copyright question worth asking first.

This is general information, not legal advice.

Google Search Scraper frequently asked questions