Google Trends Scraper (real browser)

Get Google Trends interest-over-time and related queries for up to 5 compared keywords, any region and timeframe — driven by a real headless browser (Playwright/Chromium) over an Apify residential proxy, not raw HTTP. Google has no official Trends API; a bare HTTP client hitting its unofficial JSON endpoints gets 429'd almost instantly because it has no cookies/session/consent state. This Actor loads trends.google.com like a real visitor and reads the JSON the page itself fetches for its own chart.

Why it's useful

Google Trends powers keyword research, content/SEO planning, investment sentiment signals, PR/brand monitoring and academic research — but it only has a manual web UI. This Actor turns it into a clean bulk API: compare keywords, pull the full interest-over-time series, and get the "related queries" (top + rising/breakout) for each one.

Input

{
  "keywords": ["bitcoin", "ethereum"],
  "geo": "US",
  "timeframe": "today 12-m",
  "includeRelatedQueries": true
}

• keywords — 1 to 5 search terms, compared together exactly like on trends.google.com.
• geo — two-letter country code (e.g. US, BR, GB); leave empty for worldwide.
• timeframe — Google Trends range string: now 1-H, now 7-d, today 1-m, today 3-m, today 12-m, today 5-y, or a custom YYYY-MM-DD YYYY-MM-DD range.
• category — optional Google Trends category ID (0 = all).
• hl — UI language/locale for labels, e.g. en-US, pt-BR.
• includeRelatedQueries — also fetch each keyword's related queries panel.
• proxyConfiguration — Apify Proxy for the browser session; defaults to the RESIDENTIAL group (datacenter IPs get blocked).
• timeoutSecs — how long to let the browser wait for the page's own data requests to fire (5-120, default 45).

Output (one item per keyword)

{
  "keyword": "bitcoin",
  "geo": "US",
  "timeframe": "today 12-m",
  "points": 52,
  "timeline": [
    { "time": "1727136000", "date": "2024-09-24T00:00:00.000Z", "formattedTime": "Sep 24, 2024", "value": 63, "formattedValue": "63", "hasData": true }
  ],
  "relatedQueries": {
    "top": [{ "query": "bitcoin price", "value": 100, "formattedValue": "100" }],
    "rising": [{ "query": "bitcoin etf", "value": 5000, "formattedValue": "Breakout" }]
  },
  "technique": "playwright-response-intercept",
  "scrapedAt": "2026-09-25T12:00:00.000Z"
}

How it works (technique)

1. A real Playwright/Chromium browser, over an Apify RESIDENTIAL proxy session, navigates to https://trends.google.com/trends/explore?q=...&geo=...&date=...&hl=... — exactly the URL a human comparing these keywords would open. A cookie-consent interstitial, if shown, is auto-accepted.
2. The Actor registers a page.on('response', ...) listener before navigation and watches every network response the page itself makes.
3. The page's own JavaScript fires GET /trends/api/widgetdata/multiline (the interest-over-time chart data) and GET /trends/api/widgetdata/relatedsearches (top/rising related queries per keyword) using its own session cookies — these are captured directly from the browser, never re-requested by hand.
4. Google prefixes each JSON body with )]}', (an XSSI guard) — stripped before parsing.

On blocking

Because these are the page's own requests with real cookies, they are far less likely to be blocked than the same URLs hit with a bare HTTP client (v1 of this Actor tried that and got 429'd). If the browser still can't get a chart to render within timeoutSecs (captcha wall, exhausted proxy pool, unusual keyword/geo/timeframe combo), the crawler retries with a fresh browser + proxy session up to 4 times; if every attempt fails, the Actor reports that honestly (with the real reason) instead of inventing numbers.

Use cases

• 📈 Keyword & SEO research — seasonality, trend direction, rising queries to target.
• 💹 Market/PR signal — track interest spikes around brands, tickers, products, events.
• 🎓 Research/journalism — quantify public attention to a topic over time and by region.

⭐ Enjoying this Actor?

A quick rating/review helps others find it. Want Google-Trends-style Interest-by-Region or multi-timeframe batching added? Open a ticket on the Issues tab.