Google Trends Bulk Extractor

Apify Actor that bulk-extracts Google Trends data for one or many keywords in a single run. Returns interest over time, interest by region, related queries, and (optionally) related topics, structured per keyword batch.

Why this actor

Google Trends has no public API. The unofficial google-trends-api npm package wraps Trends' internal AJAX endpoints and is the same approach most free tools use. Running it through Apify means you don't have to manage the proxy rotation, the manual URL-tweaking, or the slow per-keyword web UI — give it a keyword list, get back structured data.

Input

Field	Type	Default	Notes
keywords	string[]	required	1-5 keywords per request (Trends limit).
geo	string	""	Two-letter country code, empty = worldwide.
timeRange	string	"today 12-m"	Trends timeframe string (e.g. today 5-y, now 7-d) or custom range.
category	integer	0	Trends category ID; 0 = all.
includeInterestOverTime	boolean	true
includeInterestByRegion	boolean	true
includeRelatedQueries	boolean	true
includeRelatedTopics	boolean	false
maxRelatedItems	integer	25	Cap returned related queries/topics per rank.

If you give more than 5 keywords, the Actor automatically batches them into groups of 5 (each batch = one result item).

Output

Each dataset item represents one batch:

{
  "keywords": ["electric scooter", "e-bike"],
  "geo": "US",
  "timeframe": "today 3-m",
  "category": 0,
  "fetchedAt": "2026-07-04T20:30:00.000Z",
  "interestOverTime": { "default": { "timelineData": [...] } },
  "interestByRegion": { "default": { "geoMapData": [...] } },
  "relatedQueries": {
    "top":     [ { "query": "...", "value": 100, "link": "..." }, ... ],
    "rising":  [ { "query": "...", "value": 5000, "link": "..." }, ... ],
    "sourceShape": "rankedList",
    "default":  { ...raw upstream payload... }
  }
}

relatedQueries and relatedTopics use a normalised { top, rising, sourceShape, default } shape. The sourceShape field tells you which upstream shape the response was detected as — "rankedList" is the current google-trends-api shape (which itself is normalised through JSON round-tripping); the fallback "perKeyword" shape (older library versions / blog examples) is also handled.

If a section fails (e.g. Trends rate-limits the request or the response isn't JSON), the section is replaced by an error key like interestOverTimeError: "..." so a partial result is still useful.

Local test

npm install
npm test           # structural only - no network calls, verifies shape handling
npm run test:live  # also hits live Google Trends; rate-limits are tolerated

The structural test asserts that trimRelated correctly handles both known response shapes the google-trends-api library has shipped, that batchKeywords enforces the 5-keyword-per-batch limit, and that run() wires input/output together. The live test additionally exercises the full upstream round-trip — if Google is currently rate-limiting this IP, every section will come back with *Error keys but the test still passes (the graceful-degradation path is what we want to verify).

Deploy

This is a build-only repo. Deploy and publish to the Apify Store is done by Sky manually.

Pricing model

PAY_PER_EVENT ~$0.002/keyword-batch returned.