Google Trends Scraper

Extract Google Trends data — interest over time, interest by region, related queries, and Trending Now — via Google's internal JSON API. No login, no API key, no browser, no personal data.

What does Google Trends Scraper do?

Scrape every public Google Trends dataset in three modes:

• Keyword analysis (Explore) — compare up to 5 keywords in one query, across any country and any time window. Returns interest over time (0–100 timeline), interest by region (sub-national breakdown), and related queries (top + rising).
• Trending Now — the current trending searches for any country, with approximate traffic, publish time, and related news coverage. Uses Google's public RSS feed.
• Both — run keyword analysis and Trending Now in a single run.

Google Trends has no official public API. The announced one is a waitlisted alpha. The actor speaks to the same internal JSON endpoints the Trends website itself uses — /trends/api/explore → /trends/api/widgetdata/* — with the exact session handshake Google expects.

Rising queries are the most useful output in the whole API — they show what people are starting to search around a niche before it peaks.

Output fields

Keyword analysis (trends-query record)

Field	Description
keywords	The keyword set compared in this query
geo	Country/region code
timeRange	Time window
category	Google Trends category ID
interestOverTime	Array of { timestamp, formattedTime, values, isPartial } — one entry per timeline point, values maps each keyword to its 0–100 score
interestByRegion	Array of { geoCode, geoName, value, hasData } — sub-national breakdown
relatedQueries	Object with top[] and rising[] arrays — each entry has query, value, formattedValue, link
userConfig	Google's session classification (e.g. USER_TYPE_SCRAPER)

Trending Now (trending-search record)

Field	Description
geo	Country code
rank	Position in the trending list
keyword	Trending search term
approxTraffic	Approximate search volume (e.g. 20000+)
pubDate	When the trend was published
pictureUrl / pictureSource	Associated image
newsItems	Array of { title, url, source, snippet } — related news coverage

Who is it for?

• SEO and content teams finding keyword trends before they peak, and planning content around rising queries
• E-commerce and DTC brands reading seasonality over 5 years to plan inventory and ad budget
• Market researchers comparing the same product across countries to decide launch order
• Journalists and content creators pulling today's trending searches with approximate volume
• Brand teams tracking share of search against competitors on a normalised 0–100 scale
• Data scientists and quant analysts feeding search-interest signals into forecasting models

Pricing

$0.40 per 1,000 results. No subscription.

Results	Cost
100	$0.04
1,000	$0.40
10,000	$4.00

How to use it

1. Pick a Mode.
2. Enter Keywords (Explore), pick Trending Now Countries, or both.
3. Choose Geography and Time Range.
4. Toggle which datasets to include: interest over time, interest by region, related queries.
5. Click Start.

Output example

{
  "recordType": "trends-query",
  "keywords": ["bitcoin", "ethereum"],
  "geo": "US",
  "timeRange": "today 12-m",
  "category": 0,
  "interestOverTime": [
    { "timestamp": "1705276800", "formattedTime": "Jan 14 – 20, 2024", "values": { "bitcoin": 47, "ethereum": 38 }, "isPartial": false },
    { "timestamp": "1705881600", "formattedTime": "Jan 21 – 27, 2024", "values": { "bitcoin": 51, "ethereum": 41 }, "isPartial": false }
  ],
  "interestByRegion": [
    { "geoCode": "US-CA", "geoName": "California", "value": 100, "hasData": true },
    { "geoCode": "US-NY", "geoName": "New York", "value": 87, "hasData": true }
  ],
  "relatedQueries": {
    "top": [
      { "query": "bitcoin price", "value": 100, "formattedValue": "100", "link": "/trends/explore?q=bitcoin+price" },
      { "query": "ethereum price", "value": 62, "formattedValue": "62", "link": "" }
    ],
    "rising": [
      { "query": "bitcoin halving 2028", "value": 3200, "formattedValue": "+3,200%", "link": "" },
      { "query": "ethereum etf approval", "value": 1500, "formattedValue": "Breakout", "link": "" }
    ]
  },
  "userConfig": { "userType": "USER_TYPE_SCRAPER" },
  "scrapedAt": "2026-09-21T12:00:00.000Z"
}

Trending Now output:

{
  "recordType": "trending-search",
  "geo": "US",
  "rank": 1,
  "keyword": "hurricane watch",
  "approxTraffic": "200000+",
  "pubDate": "Mon, 21 Sep 2026 09:00:00 GMT",
  "newsItems": [
    { "title": "Hurricane watch issued for Gulf Coast", "url": "https://...", "source": "Reuters", "snippet": "..." }
  ],
  "scrapedAt": "2026-09-21T12:00:00.000Z"
}

Technical details

• No browser, no CAPTCHA, no API key. Uses Google's internal Trends JSON endpoints directly.
• Cookie-session handshake. The famous Google Trends 429 is not a rate limit — it's a missing NID cookie. The actor warms the session with a request to /trends/explore, captures the cookie even from a 429 response, and retries with it. This is the fix most broken scrapers miss.
• Anti-XSSI prefix stripping. Google prefix responses with )]}' to break naive JSON.parse calls. The actor strips it before parsing.
• Self-throttled at a configurable requestDelayMs (default 1500 ms) — Google's limiter cuts sessions at ~800 ms per request.
• Datacenter IPs work. With a valid cookie session, Google Trends accepts datacenter requests. Proxy support is available but disabled by default.
• Keyword batching. Google allows up to 5 keywords per Explore query. Larger lists are automatically split.

Known limits

• Related Topics return empty. Google marks keyless sessions as USER_TYPE_SCRAPER, and the RELATED_TOPICS widget returns [] for such sessions. The token is signed over the request — editing userType yields 401. Related Queries work fine; only Related Topics is affected.
• Session-scoped data. Google Trends returns normalised 0–100 values, not absolute search volumes. Values are comparable within a query, not across queries. Run keyword sets together to compare them.
• Google changes endpoints. The internal API is undocumented and can shift. The actor logs the exact failing URL if a call fails, making recovery fast.
• Trending Now geo coverage varies. Some countries return a rich feed, others return few or no items.
• No Google account required, and no personal data is collected — Google Trends publishes only aggregated search interest.

FAQ

Do I need an API key? No. Google Trends has no public API. The actor uses the same endpoints the Trends website uses, keyless.

Do I need a proxy? No. With a valid cookie session, Google Trends accepts datacenter IPs. Enable useProxy only if you see repeated 429s.

Why is interestByRegion empty for some keywords? Not every keyword has sub-national data. Low-volume terms are often aggregated to the country level only.

Why is relatedQueries.topics missing? Google marks keyless sessions USER_TYPE_SCRAPER and returns empty for Related Topics. Related Queries (top + rising) work fine. This is a Google-side restriction, not a bug.

How do I compare two keyword sets? Google normalises each query independently. To compare keywords, put them all in one keywords array — the actor batches them 5 at a time, and each query's 0–100 scale is internally consistent.

How do I export data? After a run, go to Storage → Export as JSON, CSV, Excel.

Support

Open an issue on the Actor's page for bugs or feature requests.