Google Trends Scraper (reliable, batch)

Fetch Google Trends data for up to 200 search terms in one run: interest-over-time plus top and rising related queries, with geo, timeframe and language control. Built for SEO pipelines, market research and AI agents — strict output schema, explicit error codes, and you only pay for terms that are actually delivered.

Why this scraper

• Only successful results are charged. A term that Google returns no data for, or that gets temporarily blocked, shows up as a clear error record at no cost.
• Anti-blocking built in. Warm session management with automatic cookie refresh, paced requests and retries on fresh connections — the failure modes that plague Trends scrapers are handled inside the actor, not left to you.
• Honest partial results. When the interest data succeeds but related queries are briefly unavailable, you get the interest data plus a note saying exactly why the related part is missing — never a silently incomplete record.
• Batch and agent-ready. Up to 200 terms per run, stable machine-readable schema.

Who this is for

SEO and content teams sizing up topics before they commit, market researchers tracking interest over time, and AI agents that need a demand signal rather than a guess.

If you check one keyword occasionally, the Google Trends website is free and fine. This actor is for pulling dozens or hundreds of terms on a schedule, in a stable format.

Who this is not for

Read this before you buy. Every term is fetched on its own, so its values are scaled to that term alone (its own peak is 100). That is what you want for tracking one topic over time, and exactly wrong for comparing two terms with each other: for that, use the Google Trends Compare Scraper from the same publisher, which puts up to five terms on one scale. Values are Google's relative index, never search volumes. For what is trending right now per country, use the Trending Now Scraper; for one term's map by country or region, the Interest by Region Scraper.

Input example

{
    "queries": ["chatgpt", "bitcoin"],
    "geo": "NL",
    "timeframe": "today 3-m",
    "includeRelatedQueries": true,
    "language": "en-US"
}

Timeframes: now 1-d, now 7-d, today 1-m, today 3-m, today 12-m, today 5-y, all. Geo: country (US, NL, DE), region (US-CA), or empty for worldwide.

Output example

{
    "query": "bitcoin",
    "status": "ok",
    "geo": "NL",
    "timeframe": "today 3-m",
    "averageValue": 56.2,
    "interestOverTime": [
        {"time": "Jun 10, 2026", "timestamp": 1781049600, "value": 90, "isPartial": false}
    ],
    "relatedQueries": {
        "top": [{"query": "koers bitcoin", "value": 100, "formattedValue": "100"}],
        "rising": [{"query": "spacex koers", "value": 950, "formattedValue": "+950%"}]
    }
}

A failed term produces an error record instead (and is not charged):

{"input": "someterm", "status": "error", "errorCode": "NO_DATA", "error": "Empty timeline"}

Error codes: BAD_QUERY, NO_DATA, BLOCKED. INPUT_TRUNCATED appears once when your input is longer than this actor accepts.

Related actors from steadydata

• google-trends-compare: up to five keywords on one scale
• google-trends-regions: where in the world the interest sits
• google-trending-searches: what is trending right now per country

Pricing

Pay per event: one trend-fetched event per delivered term - a flat price that includes the full interest-over-time series and related queries. No charge for failed terms, no separate platform-usage surcharge, and no per-row billing surprises: one term is one price, whatever the timeframe.

FAQ

Are values absolute search volumes? No — Google Trends values are relative (0-100) within your query's timeframe and region, exactly as Google reports them.

Can I compare terms? Not with this actor: each term is scaled to its own peak, so two series from this actor are not on one scale. The Google Trends Compare Scraper from the same publisher does the comparison the way the Trends Compare page does, up to five terms on one scale.

Is personal data collected? No. Trends data is aggregated by Google and contains no personal data.

Why is your price flat per term instead of per row? Because per-row billing punishes you for asking a sensible question. A twelve-month series has far more rows than a one-week series but is not worth more to you, so it costs the same here: one term, one price.

What if Google blocks a request? The actor retries on a fresh connection before giving up. If a term still fails you get a clear BLOCKED record and are not charged for it.