Google Trends has no official API, the page takes five terms at a time, and every scraper of it spends half its life on 429 errors. This is the way I know to get the Trends numbers for a whole keyword list, as a CSV, from licensed data that does not break.
Disclosure: I built the actor this uses. The data comes from Google Trends through the DataForSEO Keywords Data API under licence on every run; nothing is scraped.
What you get
One row per keyword with the full interest-over-time series as JSON, plus average_interest, latest_interest, peak_interest and peak_date, a trend_last_quarter_pct figure (the last quarter of the period against the rest) so a long list sorts without parsing the series, interest_by_region, Google's related topics and related queries (top and rising), and the google_trends_url that opens the same chart on Google.
A real row from a live run (United States, web search, past 12 months):
{
"keyword": "air fryer",
"average_interest": 66.15,
"latest_interest": 55,
"peak_interest": 100,
"peak_date": "2026-04-12",
"trend_last_quarter_pct": -15.5,
"data_points": 54
}
Top regions for that row: Wyoming 100, New Hampshire 36, South Dakota 35. Top related queries: air fryer chicken 100, air fryer oven 70, air fryer ninja 64.
What it costs
| Item | Price | 1,000 keywords |
|---|---|---|
| Keywords | $0.025 per keyword | $25.00 |
| First 3 keywords | free | −$0.08 |
| Total | about $25 |
Billing happens on your Apify account. Apify's free plan includes $5 of monthly credit, which covers a first test of about 200 keywords; a three-keyword check costs nothing.
Steps
- Open Google Trends Bulk on Apify and sign in or create a free Apify account.
- Paste your keywords into
keywords, one per line (commas work too), up to 1,000 per run. Duplicates are removed before anything is sent. - Pick
location(a country, or Worldwide),searchType(web, news, YouTube, images or Shopping) andtimeRange(7 days to since 2004). - Leave
compareoff to screen a list: every keyword is scaled on its own, so its best week is 100. Turn it on to compare up to five terms on one shared scale, exactly like adding terms on the Trends page. - Press Start. Each keyword is one request to the supplier at about 100 a minute, so 1,000 keywords take about ten minutes.
- Download CSV, Excel or JSON from the Dataset tab, or schedule it weekly with the same list. It is also an MCP tool through mcp.apify.com if an agent should call it.
Things worth knowing
- Trends values are relative, 0 to 100, never absolute search counts. To size a keyword, pair this with search volume; to see where it is heading, use the trend and peak fields here.
- A rising query value of 5000 means +5,000% against the previous period. Google's rising lists are breakout terms and often look unrelated; the top lists are the dependable ones.
- Hourly points come back for the 7-day range, daily for 30 and 90 days, weekly for 12 months and 5 years, monthly since 2004 (web search only).
- If the supplier is unreachable the run waits and retries for several minutes; if it still gets nothing, it fails with the reason and you are not charged.
If you try it and something feels off, open an issue on the actor page; I read them.
Top comments (0)