Google Trends Scraper: Interest, Regions, Related Queries

Reads Google Trends for a list of search terms and gives you, for each term:

• Interest over time, as a series of dated points.
• Interest by region, by country or by subregion.
• Related queries, both top and rising.

Up to 25 terms per run. Time ranges go from the past 7 days to 2004 to present. You can narrow by country or subregion, category and language.

It is slow on purpose. Read "How it handles Google's limits" below before you schedule it.

Who it is for

People who track the same list of terms on a schedule, not people who need thousands of terms in an hour. Some uses:

• Compare seasonal demand: when "sunscreen", "umbrella" or "flu symptoms" start rising each year.
• Watch a category: run your product terms every Monday and see which ones move.
• Feed a spreadsheet or a workflow: send each run to Google Sheets, n8n, Make or a webhook and chart it there.

If you need bulk extraction at high speed, this is the wrong tool. It will not get faster by design.

Input

Field	What it does	Default
searchTerms	The terms to read, one result per term. Up to 25 per run; extra terms are skipped.	required
geo	Country or subregion code: US, GB, BR, US-CA. Empty means worldwide.	worldwide
timeRange	now 7-d, today 1-m, today 3-m, today 12-m, today 5-y or all (2004 to present).	today 12-m
category	Google Trends category ID. 0 means all categories.	0
includeRegions	Add interest by region.	true
includeRelated	Add top and rising related queries.	true
language	Language for labels, like en-US or es-419.	en-US

Example:

{
  "searchTerms": ["cold brew", "iced tea"],
  "geo": "US",
  "timeRange": "today 12-m",
  "category": 0,
  "includeRegions": true,
  "includeRelated": true,
  "language": "en-US"
}

Duplicate terms are read once.

Output

One dataset item per term. Example, trimmed to a few points per list:

{
  "searchTerm": "cold brew",
  "geo": "US",
  "timeRange": "today 12-m",
  "category": 0,
  "interestOverTime": [
    { "date": "2026-06-14", "label": "Jun 14 - 20, 2026", "value": 100, "partial": false },
    { "date": "2026-06-21", "label": "Jun 21 - 27, 2026", "value": 97, "partial": false },
    { "date": "2026-10-18", "label": "Oct 18 - 24, 2026", "value": 58, "partial": true }
  ],
  "interestByRegion": [
    { "region": "Hawaii", "code": "US-HI", "value": 100 },
    { "region": "California", "code": "US-CA", "value": 86 }
  ],
  "relatedQueries": {
    "top": [
      { "query": "cold brew coffee", "value": "100" },
      { "query": "cold brew recipe", "value": "41" }
    ],
    "rising": [
      { "query": "cold brew concentrate ratio", "value": "+180%" },
      { "query": "protein cold brew", "value": "Breakout" }
    ]
  },
  "source": "Google Trends (unofficial reader). Values are relative, 0 to 100, not search counts.",
  "readAt": "2026-10-20T15:02:11.000Z"
}

Notes on the fields:

• The time step depends on the range and is set by Google: for example, daily points for 30 or 90 days and weekly points for 12 months.
• partial: true marks the current period, which Google is still filling in.
• relatedQueries.rising values are Google's own labels: a percentage, or "Breakout" for a very large jump.

If a term fails, its item carries only searchTerm and error, so you can see what was missed.

Price

For the launch weeks, each term read successfully costs a symbolic US$0.00001 (pay per event, event name term-result). That is a launch price, not the regular one.

After the launch the price is US$0.003 per term read successfully (US$3 per 1,000 terms). Apify notifies every user 14 days before any price increase, so the change reaches you before it applies.

A realistic case: a list of 20 terms read every week is about 86 results a month, or about US$0.26 a month at the price after the launch.

A term that fails is not charged. Apify's own platform usage for the run is billed by Apify as usual.

How it handles Google's limits

When Google answers "too many requests" (HTTP 429), the reader waits and tries again, with a longer wait each time. It does not switch IPs to get around the limit. After a few refusals on the same term, it gives up on that term and moves on. After three terms in a row fail, it stops the run instead of pushing harder.

That is why it takes about a minute per term, sometimes more. A run of 25 terms can take close to half an hour. For a job that runs once a week while you sleep, that is a fair trade: you get the data Google is willing to give, at the pace Google asks for, and you are not paying for retries that were never going to work.

What it does not do:

• No residential proxies, no IP rotation to multiply volume.
• No CAPTCHA solving, by itself or through any service.
• No Google accounts and no login.
• No fake browser fingerprints. It identifies itself with its own User-Agent that names the tool and a contact address.
• It respects Google Trends' robots.txt and only calls the paths it allows.

If Google puts a CAPTCHA or a hard block in front of Trends, the run fails with a clear message instead of trying to force its way through.

Tested from Apify on 20 October 2026: the terms it read came back complete. We have not measured a success rate over time or at volume, so we do not quote one.

Notices

• This is an unofficial tool. It is not affiliated with, endorsed by or connected to Google.
• Data source: Google Trends. Please cite it that way wherever you use the data. You are responsible for how you use it.
• Values are relative indexes from 0 to 100, not search counts. 100 is the peak for that term, place and period; 50 is half of that peak. Two runs with different settings are not on the same scale.
• Volume cap: 25 terms per run. That limit is deliberate.
• It can break if Google changes Trends. If that happens we fix it or say so on this page.
• If Google opens its official Trends API to the public, this tool moves to it. If Google or Apify asks us to take it down, we take it down the same day.
• The data contains no personal information: only the aggregate indexes Trends shows to anyone.

Who is behind it

Small Hours is run by Arlex Peralta, who answers for this tool. The work (building it, testing it and writing this page) is done by AI agents he directs. If you write to us, an AI agent may answer first, and it will say so.

Questions, bugs or requests: smallhours.tools@gmail.com, or open an issue on this page.