Google Trends Scraper & API: Compare 50 Terms, Trending Now

Google Trends data as clean, typed rows: interest over time, interest by region, related queries, and the full Trending Now list for any country. Compare up to 50 search terms on one 0–100 scale, not just 5. Built for SEO and content teams, market researchers, analysts and anyone who wants Trends data in a spreadsheet or pipeline without fighting rate limits.

Quick start

Paste this into the JSON input tab and click Start:

{"searchTerms": ["coffee", "tea", "matcha"], "geo": "US", "includeRelatedQueries": true, "trendingCountries": ["US"], "maxTrendsPerCountry": 10}

You get one row per term (12 months of weekly interest, average, peak, latest value, related queries) and today's 10 top trending searches in the US.

What you get

• Interest over time for each term: ISO dates, values 0–100, the step Google used (hour, day, week, month) and a flag on the current unfinished period.
• More than 5 terms on one scale. Google compares at most 5 terms at a time and scales each comparison on its own, so two comparisons can't be put side by side. Give this Actor up to 50 terms and it chains the comparisons through a shared anchor term and rescales them, so every term is on the same 0–100 scale and you can rank all of them.
• Interest by region: countries, regions or states, US metros or cities.
• Related queries, top and rising (with Google's "Breakout" flag), with Google's filler removed (see below).
• Trending Now: every trending search Google lists for a country in the past 4 hours, 24 hours, 48 hours or 7 days, often several hundred a day, with search volume, growth, start time, whether it is still rising, category and the searches behind it. Not just the RSS feed's newest 10.
• Only new trends for daily alerts: a schedule returns, and charges for, only trends it hasn't returned before.
• Friendly inputs: "United States" or "US", "past-90-days" or your own start and end dates, a category from a list or any of Google's ~1,400 category ids, and web, image, news, shopping or YouTube search.
• A link on every row to the same view on trends.google.com, so you can check our numbers.

Price

$2 per 1,000 rows on the Free and Starter plans ($1.50 on Scale, $1.10 on Business), plus $0.00005 per run start. You don't pay for Apify platform usage on top.

What a row is depends on the output format you pick:

Output format	One row is	Example cost
series (default)	one search term, with its whole timeline, regions and related queries inside	10 terms = 10 rows = $0.02
tidy	one date, one region or one related query	10 terms × 53 weeks = 530 rows = $1.06
Trending Now (either format)	one trending search	100 trends = $0.20

Runs that return nothing because of the input, or because nothing is new, cost nothing. Free-plan runs stop at 500 rows.

Input

Field	What it does
searchTerms	Up to 50 terms. 5 or fewer are compared directly; more are chained through an anchor term
geo	Empty for worldwide; a country name or code ("Germany", "DE"), a region ("US-CA") or a US metro ("US-NY-501")
timeRange	past-hour, past-4-hours, past-day, past-7-days, past-30-days, past-90-days, past-12-months (default), past-5-years, since-2004. Or set startDate and endDate (YYYY-MM-DD)
category, categoryId	A top-level category such as food-drink, or any Google Trends category id
searchType	web (default), images, news, shopping, youtube
includeTimeline, includeRegions, includeRelatedQueries, includeRelatedTopics	What to fetch for each term. Only the timeline is on by default
regionLevel	auto, country, region, metro (US only), city
anchorTerm	The term shared by every comparison when there are more than 5. Default: your first term
outputFormat	series (one row per term) or tidy (one row per data point)
trendingCountries, trendingWindow, trendingCategory, onlyActiveTrends, maxTrendsPerCountry	Trending Now
onlyNew	Only trends not returned by an earlier run of the same setup
maxRows	Stop after this many rows

Sample output

Twelve drinks on one scale, United States, past 5 years, weekly (a real run on 2026-10-13, outputFormat: "series"; the timeline column is left out). Google allows 5 terms per comparison, so this took three comparisons chained through the anchor coffee:

term	averageValue	peakValue	peakDate	latestValue	compared in the same request
coffee	59.89	100	2026-04-12	71	coffee, tea, matcha, kombucha, energy drink
tea	32.92	47	2026-04-12	31	coffee, tea, matcha, kombucha, energy drink
smoothie	7.62	11.09	2026-05-31	6.05	coffee, green tea, boba, smoothie
espresso	6.39	16	2026-04-12	8	coffee, hot chocolate, espresso, latte, cold brew
boba	5.07	20.17	2022-02-06	3.03	coffee, green tea, boba, smoothie
latte	3.90	7	2025-11-09	5	coffee, hot chocolate, espresso, latte, cold brew
matcha	3.68	9	2026-04-12	5	coffee, tea, matcha, kombucha, energy drink
green tea	2.84	6.05	2026-04-05	2.02	coffee, green tea, boba, smoothie
energy drink	2.45	8	2026-04-12	4	coffee, tea, matcha, kombucha, energy drink

One series row, shortened:

{
  "type": "term",
  "term": "coffee",
  "geo": "US",
  "geoName": "United States",
  "timeRange": "past-12-months",
  "resolution": "WEEK",
  "averageValue": 79.06,
  "peakValue": 100,
  "peakDate": "2026-04-12",
  "latestValue": 71,
  "latestDate": "2026-10-04",
  "timeline": [{"date": "2025-10-12", "value": 61, "isPartial": false}],
  "regions": [{"code": "US-WY", "name": "Wyoming", "value": 100}],
  "relatedQueries": {"top": [{"query": "coffee near me", "value": 100}], "rising": [{"query": "how to brew pour over coffee", "risePercent": 3350, "breakout": false}]},
  "notes": [],
  "trendsUrl": "https://trends.google.com/trends/explore?q=coffee%2Ctea%2Cmatcha&geo=US&date=today%2012-m"
}

One trending-now row (US, past 24 hours):

{
  "type": "trending-now",
  "country": "US",
  "title": "harrison mevis",
  "searchVolume": 1000000,
  "searchVolumeText": "1,000,000+",
  "increasePercent": 1000,
  "startedAt": "2026-10-13T00:30:00Z",
  "endedAt": null,
  "active": true,
  "categories": ["Sports"],
  "relatedQueries": ["rams kicker", "bills", "buffalo bills"],
  "window": "past-24-hours",
  "source": "trending-now"
}

With outputFormat: "tidy" the same data comes as interest-over-time rows (term, date, value, isPartial, rawValue), interest-by-region rows (regionCode, regionName, value) and related-query rows (list, rank, query, value, risePercent, breakout). The dataset has a view for each row type, and the run's SUMMARY record lists the comparisons made, the anchor, any filler removed and anything that failed.

How the 0–100 numbers work

Google never publishes search counts. Each value is search interest relative to the highest point in what you asked for: 100 is the peak, 50 is half as popular, 0 means too little data. Region maps are on each term's own scale (100 = the region where that term is most popular). Values from different runs are on different scales; compare terms inside one run.

With more than 5 terms, the anchor term appears in every comparison. The ratio of the anchor's totals between two comparisons converts one scale into the other, and the result is renormalised so the highest point of all terms is 100. rawValue keeps Google's own number. Google rounds each comparison to whole numbers relative to its biggest term, so a small term that shares a comparison with a giant one gets coarse values (kombucha reads 1 to 2 next to coffee above). The anchor sits in every comparison, so a giant anchor flattens all of them: pick a term of middling popularity as anchorTerm (here espresso rather than coffee) and the small terms keep their detail. If the anchor reads 0 to 9 in a comparison, that row's scaleNote says its conversion is approximate.

Honest limits

• Google rate-limits Trends hard. This Actor spaces requests, backs off on "too many requests", rotates IPs and moves to residential proxies only when it must. A very large request can still run out of time; it then stops with what it has and a message saying what to narrow, instead of failing.
• Filler in related queries. Google labels API clients like this one as automated and, for some terms, pads related lists with stock phrases ("gardening", "hospitals", "how to cook pasta" for "tea"). By default clusters of those are removed and listed in the run's SUMMARY record. Turn off filterNoise to get Google's lists exactly as returned. A real but generic related query can occasionally be removed with them.
• Related topics currently come back empty from Google for most terms when requested by an automated tool. The option is there and the run tells you when Google returned nothing.
• Trending Now search volumes are Google's own lower bounds ("20K+" becomes 20000). If the full list is unavailable, the Actor falls back to the public RSS feed (newest 10 per country, with news headlines) and says so.
• Data comes from Google's public Trends pages. Cite it as "Data source: Google Trends (https://www.google.com/trends)" when you publish it, as Google asks.

FAQ

Can I compare more than 5 terms? Yes, up to 50 in one run, on one scale. See "How the 0–100 numbers work".

Why do my numbers differ slightly from trends.google.com? Google samples searches, so the same query can return slightly different values from one request to the next. Use the trendsUrl on each row to compare.

Can I schedule daily trend alerts? Yes. Set trendingCountries, turn on onlyNew, and schedule the Actor. Each run returns only trends it hasn't returned before.

Is there a free trial? Apify's free plan includes monthly credit that covers thousands of rows here.