Google Trends Scraper: interest over time, by region, and related queries

Type up to 25 keywords and get Google Trends back as flat rows: one per point on the interest chart, one per region, one per related search, and one per entry in a country's Trending Now list. No Google account and no API key.

Read this before you build on it. Every value is Google's own 0 to 100 index, not a search volume. 100 is the peak of the exact series you asked for, so changing the region, the time range or the category rescales the whole thing, and two keywords pulled in separate runs are not comparable. Google does not publish absolute search counts to anyone, so no tool can give you them.

Input	Up to 25 keywords, and country codes for Trending Now
Output	One row per datapoint, region, related query or trending search
Ceiling	20,000 rows per run
Account needed	None, and no Google API key
Price	$0.06 per 1,000 rows, flat on every plan

🔍 What Google Trends Scraper does

It pulls the three panels the Trends website shows for a keyword, and flattens each into rows.

Interest over time is one row per point on the chart. The range decides how fine the points are: the past day is minute by minute, a year is weekly, all is monthly back to 2004.

Interest by region is one row per place. Worldwide runs break down by country, a single country by state or province, or by city or metro area if you ask.

Related and rising queries is one row per related search, marked top for the ones consistently searched alongside your keyword and rising for the ones growing fastest.

Trending Now is separate. Give it country codes and you get what is spiking in each one now, with the top news story behind it. No keywords needed.

📥 What you give it

{
  "searchTerms": ["bitcoin", "ethereum"],
  "geo": "US",
  "timeframe": "today 12-m",
  "dataTypes": ["interestOverTime", "interestByRegion", "relatedQueries"],
  "maxItems": 500
}

Field	Default	What it is
searchTerms	box starts at bitcoin, ethereum	Keywords or phrases, up to 25. Each is looked up on its own.
geo	box starts at US	A two-letter country, a sub-region like US-CA or GB-ENG, or empty for worldwide.
timeframe	box starts at today 12-m	One of the ranges in the dropdown, or a custom range typed as 2024-01-01 2024-12-31.
dataTypes	all three ticked	interestOverTime, interestByRegion, relatedQueries. Untick what you do not need and the row count falls with it.
maxItems	box starts at 1000	Total rows for the run, up to 20,000.
compareKeywords	off	Puts the first five keywords on one shared scale. Needs at least two keywords to do anything.
searchType	web	web, images, news, youtube or shopping.
regionLevel	default	country, region, city or metro. City and metro need a single country, and metro is United States only.
category	0	Google's own numeric category id, like 71 for Food and Drink. 0 is every category.
language	en-US	How region names and labels are spelled. It changes no numbers.
trendingNowGeos	none	Country codes for the Trending Now list, up to 20. Usable on its own.
proxyUrls	none	Optional. Your own servers, used exactly as given.

Budget about 150 rows per keyword with all three data sets over twelve months. City and metro detail add roughly 200 rows each on their own.

A geo the run cannot read is treated as worldwide, not as an error. It wants US, not USA or United States. A timeframe it cannot read falls back to the past twelve months the same way, so check the first run's rows match the window you meant.

📤 What you get back

A real row from a recent run:

{
  "ok": true,
  "keyword": "bitcoin",
  "geo": "US",
  "timeframe": "today 12-m",
  "searchType": "web",
  "category": 0,
  "trendsUrl": "https://trends.google.com/trends/explore?q=bitcoin&date=today+12-m&geo=US",
  "recordType": "interest_over_time",
  "date": "2025-09-21T00:00:00.000Z",
  "dateLabel": "Sep 21, 2025",
  "value": 26,
  "formattedValue": "26",
  "isPartial": false,
  "scrapedAt": "2026-09-21T01:23:40.241Z"
}

Field	What it is
recordType	interest_over_time, interest_by_region, related_query or trending_now. Split the dataset on it.
value	Google's 0 to 100 index for this row. Not a volume.
formattedValue	Google's own label. On rising related queries this is where Breakout and +450% appear, which the number cannot express.
date	ISO 8601 in UTC, on interest-over-time rows and on a trending row's publication time. null on regional and related rows.
dateLabel	The same instant as Google prints it on the chart axis, kept because a weekly point covers a range the timestamp hides.
isPartial	True on a final point whose period has not finished. It will move. Drop those before charting a trend line.
geoName, geoCode	The region name and its code, on regional rows. latitude and longitude appear instead at city level.
relatedQuery, queryType, rank	The related search, top or rising, and its position with 1 strongest.
trafficEstimate, newsTitle, newsUrl, newsSource	Trending Now rows only. The approximate search count Google publishes, and the story behind it.
keyword, geo, timeframe, searchType, category	Repeated on every row on purpose, so you can filter or group without joining anything back.

🧾 Reading the output

Three kinds of row can land in your dataset.

Row	How to spot it	Charged
A datapoint	_sample and _diagnostic both absent	yes
The sample row	_sample: true and recordType: "sample"	no
A diagnostic	_diagnostic: true and an errorCode	no

An empty input writes the sample row and stops there, so you can see the shape before spending anything.

Code	What it means
NO_RESULTS	Google has no data for that keyword in that region and range. Broaden one of them.
RATE_LIMITED	Google throttled that request and would not answer it. Re-run, or ask for less in one go.
BLOCKED	Google refused the request.
NOT_FOUND	Google's trending feed for that country did not answer.
NETWORK	Google was unreachable or answered badly.
TIME_BUDGET	The run hit its time limit before reaching that keyword. Everything already delivered is complete.
PROXY_INPUT_ADJUSTED	A network setting in your input was not usable, so the run used its own.
CHARGE_ERROR	A charge could not be recorded. The run stops handing over rows rather than continuing quietly.

One keyword with no data never stops the others, and a diagnostic row never fails the run.

▶️ How to run it

1. Open Google Trends Scraper and click Try for free.
2. Put your keywords into Keywords, deleting the examples first.
3. Set Region and Time range, and untick anything in What to return you do not need.
4. Set Maximum rows. Around 200 is enough to see what one keyword produces. Then click Start.
5. Download the dataset as JSON, CSV or Excel, or read it from the Apify API.

💰 How much does it cost?

$0.06 per 1,000 rows. Flat on every Apify plan, no volume tiers.

Every delivered row counts the same, whichever panel it came from. Points Google marks as having no data are dropped before they become rows, so a thin series costs nothing for its empty part. Sample rows, diagnostic rows, a keyword with no data and a region detail level Google does not publish are all free.

💡 What people use it for

• Checking whether a product name is actually growing before committing a budget to it.
• Pulling rising related queries every week and using them as the content or keyword backlog.
• Finding which states or cities index highest for a term, then pointing spend at those first.
• Watching the Trending Now list for a country on a schedule, with the news story attached.
• Putting up to five names on one shared scale with compareKeywords to settle an argument.

🚧 What it does not do

• No absolute search volumes. The 0 to 100 index is all Google publishes.
• No related topics. Google returns an empty list for that panel to callers who are not signed in, so shipping the field would mean shipping an always-empty column. Related queries are unaffected.
• Two separate runs are not comparable. Each keyword is scaled against its own peak unless you turn on compareKeywords and run them together.
• Comparison mode takes five keywords, which is Google's own limit. Extras are dropped and the run says so in the log.
• A low-volume keyword returns nothing, not a flat line of zeroes. Google suppresses series it considers too thin, and you get a free NO_RESULTS row.
• City and metro detail is best effort. Google publishes dense city data for big terms in big countries and nothing at all for many other combinations, and the run falls back to the level it does publish.
• The last point of a series is usually partial and will move if you re-run later the same day.
• Trending Now is a live snapshot, typically around ten entries, and two runs an hour apart will differ.
• Category ids are Google's numbers, with no lookup built in. Read the id off a Trends URL after narrowing a search there.

🧭 Which Google scraper do you need?

If you want	Use
Interest in a keyword over time, by region, and what else people search	This one
The actual results pages for those keywords	Google Search Results Scraper
News coverage of a spiking topic	Google News Scraper
App listings for a name that is trending	Google Play Apps Scraper
Places and reviews rather than searches	Google Maps Scraper

❓ Questions people ask

Do I need a Google account or an API key? No. Nothing to sign up for and no quota.

Why does one keyword produce so many rows? Because the output is flat. Twelve months of weekly points, a row per region and a row per related query is about 150 rows for a single keyword. Untick data sets in dataTypes or lower maxItems for fewer.

Can I compare two keywords directly? Turn on compareKeywords and put both in searchTerms. Up to five then share one 0 to 100 scale.

What happens when a keyword has no data? One free NO_RESULTS row for it, and the run carries on to the rest.

Can I run it on a schedule? Yes. Nothing carries over between runs. Key on keyword, recordType and date, and remember the last point is still moving.

Is this legal? Trends is public, aggregated and carries no personal data. Apify's write-up on scraping and the law is a good starting point, and we are not lawyers.

🆘 If something breaks

Open the Issues tab on the actor page. Send the keyword, the region and the run ID. The errorCode on the diagnostic row usually names the problem on its own.