Google Trends Scraper — Fast, No Browser Needed

Scrape Google Trends data via Google's own internal JSON API — no browser, no Playwright, no CAPTCHAs. Get interest over time, interest by region, related queries and related topics for any keyword, geo and time range. A fast alternative to the official Google Trends UI and other browser-based scrapers.

Why this Actor

• Fast — hits Google's JSON API directly instead of rendering a browser. A typical keyword completes in seconds.
• Reliable — automatic retries with exponential backoff on rate limits (HTTP 429), a real cookie jar (NID, AEC, CONSENT, SOCS) that survives Google's EU consent redirects, and you're never charged for empty results. Optional Apify Proxy to escape IP soft-blocks.
• Clean output — structured records with typed values, ready for Excel, Looker Studio, or an LLM pipeline.

What you get

Per keyword (configurable via Data types):

Data type	Description
INTEREST_OVER_TIME	Search interest over the selected time range (0–100 scale)
INTEREST_BY_REGION	Interest broken down by country / region
RELATED_QUERIES	Top and rising related search queries with interest values
RELATED_TOPICS	Top and rising related topics

Input

Field	Type	Description
searchTerms	array	Keywords to analyze. Multiple keywords = comparison in one run.
startUrls	array	Optional — paste Google Trends explore URLs (keyword, geo, date parsed automatically). Comparison URLs with comma-separated terms (?q=react,vue) are split into individual keywords. Overrides searchTerms when provided.
geo	string	ISO country code (e.g. US, LT, DE, GB). Empty = Worldwide.
timeRange	select	now 1-H … all (since 2004). Default: today 12-m.
customTimeRange	string	Exact range as YYYY-MM-DD YYYY-MM-DD (takes precedence).
category	integer	Google Trends category ID (0 = all).
dataTypes	multi-select	Which datasets to return per keyword.
maxItems	integer	Cap on output records (0 = unlimited).

Example input

{
  "searchTerms": ["chatgpt", "claude"],
  "geo": "US",
  "timeRange": "today 3-m",
  "dataTypes": ["INTEREST_OVER_TIME", "RELATED_QUERIES"]
}

Output

One record per (keyword × data type). Example:

{
  "type": "INTEREST_OVER_TIME",
  "keyword": "chatgpt",
  "geo": "US",
  "timeRange": "today 3-m",
  "comparisonGroup": "US|today 3-m|chunk-0",
  "data": [
    { "date": "May 7, 2026", "value": 95 },
    { "date": "May 14, 2026", "value": 92 }
  ]
}

{
  "type": "RELATED_QUERIES",
  "keyword": "chatgpt",
  "geo": "US",
  "timeRange": "today 3-m",
  "comparisonGroup": "US|today 3-m|related:chatgpt",
  "top": [{ "query": "openai chatgpt", "value": 100, "formattedValue": "100", "link": "/trends/explore?q=..." }],
  "rising": [{ "query": "chatgpt wrapper", "value": null, "formattedValue": "Breakout", "link": "/trends/explore?q=..." }]
}

comparisonGroup identifies the explore request a record came from. Google normalizes interest to 0–100 per request, so only records with the same comparisonGroup are mutually comparable.

Export to CSV / JSON / XLSX from the dataset, schedule recurring runs, or call it via API for programmatic access.

Use cases

• SEO & content strategy — find rising queries before they peak
• Market research — compare brand interest across regions
• Trend monitoring — scheduled runs to track momentum over time
• AI/LLM pipelines — structured trend data as model input

Cost

Pay-per-result: $1.50 per 1,000 dataset records, plus $0.01 per Actor start. You are charged only for records pushed to the dataset — never for empty results. Run a small test first — a single keyword with all data types costs cents.

Notes

• Interest values are Google's normalized 0–100 scale.
• More than 5 terms? Values across chunks are NOT comparable. Google caps comparisons at 5 terms per explore, so larger groups are split into multiple requests — each normalized to 0–100 independently. A 95 in chunk-0 and a 60 in chunk-1 are on different scales. Only compare records that share the same comparisonGroup.
• RELATED_TOPICS may return an empty list for some keyword/geo combinations — Google doesn't always have topic data.
• The Actor uses Google's public Trends JSON endpoints. Usage falls under Google's Terms of Service for the Trends service.

Development

npm install
apify run --input-file /tmp/input.json   # test locally
apify push                                 # deploy to Apify