Google Trends Scraper — Reliable

Google Trends data that actually arrives.

If you have used Google Trends tools before, you know the pattern: the run starts, Google rate-limits it, and you get an error — or worse, an empty dataset — while still paying for the compute. This actor is built specifically to not do that. It treats reliability as the product:

• Proper session handling. Google Trends serves data only to a correctly bootstrapped session (cookies first, then the request token). The actor builds that session the way a browser does, and rebuilds it automatically if Google invalidates it mid-run.
• Pacing and retries with backoff. Requests are spaced out, and any rate-limit (HTTP 429), forbidden (403), or server error is retried with growing waits — 5s, 15s, 30s — instead of failing on the first hit.
• Partial results instead of dead runs. If one section still cannot be fetched after every retry, the run delivers everything else and tells you exactly which section is missing (see the run_summary item). You never pay for an all-or-nothing failure.
• Optional Apify Proxy support for large or repeated runs, which keeps success rates high when Google gets aggressive.

What you get

For 1–5 keywords, in any country (or worldwide), over preset time ranges from the past hour to the past 5 years:

• Interest over time — all keywords compared on one shared 0–100 timeline, like Google Trends itself.
• Interest by region — per keyword, which regions search most.
• Related queries — the top and rising searches related to each keyword.
• Related topics — the top and rising topics related to each keyword. (Note: Google sometimes serves no topic rows for a query. When that happens the section returns empty top/rising lists rather than failing — the data simply is not there.)
• Trending searches (optional) — what is trending now in the chosen location, with approximate traffic and the headlines driving it.

Pricing

Pay per result. You are charged one low per-result fee for each dataset item the actor delivers: each interest-over-time item, each by-region item, each related-queries item, each related-topics item, and the trending-searches item. The final run_summary item is always free — you never pay for the bookkeeping. There is no subscription and no charge for a section that failed: if it is not in your dataset, you did not pay for it.

Apify platform usage (compute) is billed by Apify separately, as with any actor. Enabling Apify Proxy adds Apify's proxy charges.

Input

Field	Type	Default	Description
keywords	array of strings (1–5)	["coffee"]	Search terms to look up. Multiple terms are compared on one shared scale.
geo	string	US	Country code, e.g. US, GB, DE; region codes such as US-CA work too. Use Worldwide (or empty) for worldwide results.
timeframe	select	Past 3 months	Past hour, past 4 hours, past day, past 7 days, past month, past 3 months, past 12 months, or past 5 years.
category	integer	0	Google Trends category ID. 0 = all categories.
sections	checkboxes	all four	Interest over time, interest by region, related queries, related topics.
includeTrending	boolean	true	Add a trending-searches item for the location (worldwide runs fall back to US trending).
useProxy	boolean	false	Route requests through Apify Proxy. Recommended for large or repeated runs.

Example input

{
  "keywords": ["coffee", "tea"],
  "geo": "US",
  "timeframe": "today 3-m",
  "category": 0,
  "sections": ["timeseries", "geo", "related_queries", "related_topics"],
  "includeTrending": true,
  "useProxy": false
}

Output

Every run writes these item types to the default dataset:

timeseries — one item per run

{
  "type": "timeseries",
  "keywords": ["coffee", "tea"],
  "geo": "US",
  "timeframe": "today 3-m",
  "points": [
    { "date": "2026-07-31", "timestamp": 1785456000, "values": { "coffee": 74, "tea": 31 } }
  ]
}

values maps each keyword to its 0–100 interest index for that date, so compared keywords share one scale.

geo — one item per keyword

{
  "type": "geo",
  "keyword": "coffee",
  "geo": "US",
  "timeframe": "today 3-m",
  "regions": [ { "region": "Hawaii", "value": 100 } ]
}

related_queries — one item per keyword

{
  "type": "related_queries",
  "keyword": "coffee",
  "geo": "US",
  "timeframe": "today 3-m",
  "top": [ { "query": "coffee near me", "value": 100 } ],
  "rising": [ { "query": "coffee protein", "value": 4350 } ]
}

Rising values are percentage growth; very large growth may appear as a breakout value in the tens of thousands or higher.

related_topics — one item per keyword

{
  "type": "related_topics",
  "keyword": "coffee",
  "geo": "US",
  "timeframe": "today 3-m",
  "top": [ { "title": "Coffee", "type": "Beverage", "value": 100 } ],
  "rising": []
}

As noted above, Google sometimes serves no topic rows; the lists are then empty and the section still counts as delivered.

trending — one item per run (when enabled)

{
  "type": "trending",
  "geo": "US",
  "items": [
    { "title": "some headline topic", "approx_traffic": "500K+", "published": "2026-10-28T04:00:00+00:00", "news_titles": ["..."] }
  ]
}

run_summary — one item per run, always last, always FREE

{
  "type": "run_summary",
  "succeeded": ["timeseries", "geo", "related_queries", "related_topics", "trending"],
  "failed": [],
  "duration_s": 24.61
}

succeeded and failed list section names. If a section is in failed, its items are absent from the dataset and you were not charged for them — everything else in the run is complete and usable.

Tips

• Start with the defaults. Turn on Use Apify Proxy when you run many keywords, short timeframes (past hour/day), or scheduled frequent runs.
• Comparing keywords (2–5 at once) is the cheapest way to get relative interest: one timeseries item covers all of them on a shared scale.
• For worldwide runs, the trending block falls back to US trending searches, since there is no single worldwide trending feed.