Google Trends Scraper: Interest, Related Queries, Trending

Get Google Trends data with one result per keyword. Each keyword's result holds all of this:

• Interest over time for the keyword, alone or compared with up to 4 others (0-100 scale, as in the Trends UI)
• Interest by region (states/regions for a country, countries for worldwide)
• Related queries, both top and rising (including "Breakout")

It can also list trending now: Google's current trending searches per country, with approximate traffic, picture and news links, one result per search.

It reads Google Trends' own JSON and RSS endpoints over plain HTTP. There is no browser, so runs are light and quick.

Input

Field	Description
keywords	Search terms. Every 5 keywords form one comparison group (like adding terms in the Trends UI), so values in a group are relative to each other.
geo	Country or region code (US, GB, US-CA). Empty = worldwide. Default US.
timeframe	now 1-H, now 4-H, now 1-d, now 7-d, today 1-m, today 3-m, today 12-m, today 5-y, all, or YYYY-MM-DD YYYY-MM-DD. Default today 12-m.
category	Google Trends category ID. 0 = all.
property	"" web, news, images, youtube, froogle (Shopping).
dataTypes	Any of interestOverTime, interestByRegion, relatedQueries. Default all three.
trendingNowGeos	Optional country codes for trending searches, e.g. ["US", "GB"].
maxItems	Cap on charged results (one per keyword, one per trending search). Keyword results come first.

You can run only trending searches by leaving keywords empty and setting trendingNowGeos.

Output

One row per keyword, plus one row per trending search. type is keyword or trendingNow.

keyword

{"type": "keyword", "keyword": "bitcoin", "comparedWith": ["ethereum"], "geo": "US", "timeframe": "today 12-m",
 "category": 0, "property": "",
 "interestOverTime": [{"date": "2025-11-02T00:00:00+00:00", "timestamp": 1762041600, "value": 31, "isPartial": false}, ...],
 "interestByRegion": [{"geoName": "Wyoming", "geoCode": "US-WY", "value": 81}, ...],
 "relatedQueries": {"top": [{"query": "bitcoin price", "value": 100, "formattedValue": "100", "link": "https://trends.google.com/..."}, ...],
                    "rising": [...]},
 "missing": [], "scrapedAt": "2026-11-01T06:27:07+00:00"}

• Keywords in one group of up to 5 are compared, as in the Trends UI. interestOverTime values are on one 0-100 scale across the group. comparedWith lists the other keywords.
• In a compared group, interestByRegion values are each keyword's share of the group's searches in that region, so they add up to 100 across the group (Wyoming: bitcoin 81, ethereum 19). For a single keyword they are the usual 0-100 interest score.
• Regions Google has no data for are left out. isPartial is true for the latest, still incomplete period.
• missing lists data types Google didn't return after retries.
• A keyword with no data at all (nobody searches it, or every request failed) comes back with an error and is not charged.

trendingNow: geo, title, approxTraffic (e.g. 2000+), pubDate, picture, pictureSource, newsItems (list of title, url, source, picture).

Fields Google doesn't send are null.

Reliability

• The Actor first loads trends.google.com to get the session cookie Google expects.
• On HTTP 429/403 it waits 2, 4, then 8 seconds, refreshes the cookie once and retries.
• Keyword groups are requested about a second apart.
• The log ends with the count of HTTP 200 and 403/429 responses.
• If every keyword request fails, the run fails instead of returning only trending rows.
• A failed related-queries or region request is retried, and if it still fails the keyword's missing field says so.

Related topics are not included: Google returns them empty to this kind of request.

Pricing

Pay per result: one result is one keyword's full report, or one trending search. Keywords with no data are free.