Google Trends Scraper

Get Google Trends data for many keywords and countries in one run: interest over time, interest by region, related queries (top and rising) and today's trending searches. Output is clean JSON, CSV or Excel.

Google rate-limits Trends hard (the famous 429 errors). This actor retries every single request from a fresh IP, so a run finishes with complete data instead of half-empty rows. You pay only for terms that come back with data.

Typical uses: keyword and SEO research, content planning, seasonality and demand forecasting, product and market research by country or US state, brand monitoring, spotting breakout queries.

What you get

Field	Notes
keyword, geo, timeRange	One row per search term and country (WORLD for worldwide)
status	ok, no_data (Google has no data for the term), partial (timeline fine but regions or related queries could not be fetched) or failed. Only ok rows are charged
average, peak	Interest on Google's 0-100 scale; peak has the date and value
timeline[]	Every data point: date, label, value, plus isPartial / noData flags
regions[]	Countries, states, cities or metro areas where the term is searched most (geoCode, geoName, value)
relatedQueries.top[] / .rising[]	Related searches with Google's score or Breakout / +N% growth, and a link to each
comparedWith	Other terms in the same comparison group
trendsUrl	Link to the same chart on trends.google.com

Trending mode returns Google's daily trending searches per country: rank, term, approxTraffic, publishedAt and the news stories behind each trend.

How to use

1. Add Search terms and one or more Countries / regions (US, DE, TR, or a subregion such as US-CA; WORLD for worldwide).
2. Pick a Time range: from the past hour (minute-level points) to 2004-present, or a custom YYYY-MM-DD YYYY-MM-DD range.
3. Turn on Compare terms when you want values that are directly comparable. Google scales every chart to 100 for its highest point, so terms are only comparable when they are requested together (groups of up to 5).
4. Run it and download the dataset.

Other filters: Google property (web, news, YouTube, images, shopping), category ID, region detail (countries, states, cities, US metro areas) and interface language.

Output example

{
  "type": "interest",
  "keyword": "rust",
  "geo": "US",
  "timeRange": "today 12-m",
  "status": "ok",
  "average": 21,
  "peak": { "date": "2026-03-08T00:00:00.000Z", "value": 100 },
  "comparedWith": ["python"],
  "timeline": [{ "date": "2025-10-12T00:00:00.000Z", "label": "Oct 12 – 18, 2025", "value": 18 }],
  "regions": [{ "geoCode": "US-WY", "geoName": "Wyoming", "value": 100 }],
  "relatedQueries": {
    "top": [{ "query": "what is rust", "value": 100, "label": "100", "url": "https://trends.google.com/trends/explore?q=what+is+rust..." }],
    "rising": [{ "query": "rust 1.80", "value": 5000, "label": "Breakout", "url": "https://trends.google.com/trends/explore?q=rust+1.80..." }]
  },
  "trendsUrl": "https://trends.google.com/trends/explore?q=python%2Crust&date=today+12-m&hl=en-US&geo=US",
  "scrapedAt": "2026-10-13T12:00:00.000Z"
}

Pricing

Pay per event: a small start fee per run (charged per GB of memory), a fee per search term result and a smaller fee per trending search. Terms Google has no data for, partial results and failed lookups are free. Proxy costs are included. Higher Apify plans get a lower price per result. Use Max results to cap spending.

Good to know

• Google Trends shows relative interest (0-100), not search volumes. Values from different runs or different groups are not comparable unless the terms were compared together.
• Very specific or low-volume terms often return no data; these rows get status: no_data and are not charged.
• Related topics are not offered: Google currently returns them empty.
• Region detail depends on Google: US metro areas (DMA) and states work well, while city-level data is sparse and worldwide runs return mostly countries.
• Hourly and daily ranges do not include Google's own average, so average is computed from the returned points.
• Terms in a group of 5 are comparable only with each other; with more than 5 terms, each group of 5 is scaled separately.
• Trending mode reads Google's public trending feed, which lists about 10-25 searches per country at a time.
• If Google blocks a request the actor retries it from a new IP. Terms that still fail after several rounds are listed in the SUMMARY record of the run and returned as failed rows.
• Concurrency defaults to 4. Raise it for speed, lower it if you see many retries in the log.
• Only public aggregate data is collected; no personal data. Use the data in line with Google's terms and the laws that apply to you.