Google Trends Scraper Pro: trends by term and region

Scrape public Google Trends data by term, geo and time range: interest over time, interest by region and city, related queries and related topics. The actor uses HTTP requests without a browser or login. Google may rate-limit requests, so residential proxies are recommended for production runs.

What it does

Google Trends Scraper Pro extracts, for 1 to 20 search terms per run:

• Interest over time — daily/weekly time series normalized 0–100, plus the period average.
• Interest by region — countries worldwide, or states/subregions when you set a country geo.
• Interest by city — city-level interest when a country geo is selected.
• Related queries — TOP and RISING (including Breakout queries with their growth rate).
• Related topics — TOP and RISING, when Google returns mapped topics for the term.

One dataset item per term. A term with no data is saved with hasData: false instead of failing the whole run.

Why it is more reliable (pure HTTP, no browser)

This actor calls the JSON endpoints used by the Trends frontend (/trends/api/explore and /trends/api/widgetdata/*):

1. Prime session cookies (cookieless API calls are rejected).
2. GET /trends/api/explore with the comparison item → widget tokens.
3. GET /trends/api/widgetdata/multiline (time series), comparedgeo (regions, cities) and relatedsearches (top + rising).
4. Normalize into one item per term.

Resilience is built in: 3 attempts with exponential backoff on 429/5xx/timeouts, session cookie re-priming between retries, rotating residential proxy support, polite spacing between calls, and per-item error isolation — one bad term never kills the run.

Use cases

• SEO & Keyword Strategy: Discover breakout search queries and seasonal search spikes before competitors do.
• Competitor Benchmarking: Compare up to 20 brand or product keywords on an aligned 0–100 scale to gauge relative search interest.
• Geo-Targeted Marketing: Identify which countries, states, or cities have the highest concentration of search volume for precise ad spend allocation.
• Content & Trend Forecasting: Track rising topics and queries over 7 days, 3 months, or 5 years to plan editorial calendars and viral campaigns.
• AI Agent & LLM Tooling: Feed real-time aggregate trend data directly into AI assistants and automated research workflows via Apify and MCP.

Input

Field	Type	Description
searchTerms	string array (required)	Terms to query, 1–20 per run.
timeRange	string	today 3-m (default), today 12-m, today 5-y, all, now 7-d, or a custom YYYY-MM-DD YYYY-MM-DD range.
geo	string	Country code (US, AR, ES) or empty for worldwide.
category	integer	Numeric Google Trends category (0 = all).
maxItems	integer	Max results to save (default: 5, 0 = no limit). Caps your spend.
proxyConfiguration	proxy	Residential rotating proxies recommended. Works without proxy too.

Example input (JSON)

{
    "searchTerms": ["web scraping"],
    "timeRange": "today 3-m",
    "maxItems": 5
}

You can pass up to 20 terms sharing the same timeRange/geo. Google compares at most five terms per request. For larger sets, the actor joins groups using a repeated reference term and rescales the timelines to one 0–100 scale. Each item carries a comparison table with every term's average on that scale. If Google does not return enough reference data to align the groups, the actor falls back to per-term series and comparison is null.

Output example (real run)

{
    "searchTerm": "web scraping",
    "timeRange": "today 3-m",
    "geo": "",
    "timeline": [
        {
            "time": "1783209600",
            "formattedTime": "Jul 5, 2026",
            "value": 37
        },
        {
            "time": "1783296000",
            "formattedTime": "Jul 6, 2026",
            "value": 33
        },
        {
            "time": "1783382400",
            "formattedTime": "Jul 7, 2026",
            "value": 33
        }
    ],
    "averages": [
        27.8
    ],
    "interestByRegion": [
        {
            "geoCode": "SH",
            "geoName": "St. Helena",
            "value": 100
        },
        {
            "geoCode": "CN",
            "geoName": "China",
            "value": 30
        },
        {
            "geoCode": "NL",
            "geoName": "Netherlands",
            "value": 29
        }
    ],
    "interestByCity": [],
    "relatedQueriesTop": [
        {
            "query": "python web scraping",
            "value": 100,
            "formattedValue": "100"
        },
        {
            "query": "web scraping ai",
            "value": 94,
            "formattedValue": "94"
        },
        {
            "query": "what is web scraping",
            "value": 72,
            "formattedValue": "72"
        }
    ],
    "relatedQueriesRising": [
        {
            "query": "web scraping là gì",
            "value": 200,
            "formattedValue": "+200%"
        }
    ],
    "relatedTopicsTop": [],
    "relatedTopicsRising": [],
    "comparison": null,
    "hasData": true
}

Notes: interestByCity is populated when a country geo is set and Google reports city-level data (low-volume terms legitimately return none). relatedTopics* mirrors what Google returns — terms without mapped knowledge graph entities return empty arrays. Empty arrays mean "Google returned nothing", never invented data. comparison is null on single-term runs; on multi-term runs it lists every term with its average on the shared 0–100 scale, e.g. "comparison": [{"searchTerm": "web scraping", "average": 62.4}, {"searchTerm": "data extraction", "average": 31.8}].

Pricing

Pay per event, charged only for results actually saved:

Event	Price
result-item — one processed term	US$ 0.002

Example: 1,000 results ≈ US$ 2.00. maxItems caps the run, and the actor also reads your max-charge budget before starting, so a run never spends more than you allowed.

Comparison with alternatives

This actor calls the public JSON endpoints directly and charges $0.002 per saved result. Google can still rate-limit datacenter endpoints; an Apify residential proxy is recommended in production. Each result includes the fields Google returns for that term: period averages, geographic interest, and TOP and RISING related queries and topics when available.

FAQ

Is there an official Google Trends API? No. Google publishes no official API; this actor uses the same public JSON endpoints the trends.google.com frontend calls. Only public, aggregate trend data is collected.

What data does it collect? Public aggregate trend statistics. It does not request account credentials or personal search histories.

Why is interestByCity empty? City resolution only exists for country geos, and Google reports cities only where search volume is significant. Worldwide runs and niche terms legitimately return [].

Why is timeline empty / hasData: false? Google returned no measurable interest for that range/geo/category. Widen the range or drop the geo filter. A request failure is logged and skipped rather than reported as a no-data result.

I get many 429s. Google throttles datacenter IPs. Enable Apify Proxy (residential) in the input — that is exactly what the proxyConfiguration field is for.

How are values normalized? 0–100 relative to the peak in the requested region and period (100 = peak popularity). On multi-term runs the actor aligns groups of at most five terms to one shared scale when Google supplies overlapping reference data; comparison summarizes each term's average on that scale.

Integrations

Results land in the default dataset: pull them via the Apify API (/v2/actor-runs/{runId}/dataset/items), export CSV/JSON, or push them to Make, Zapier, or any webhook. The actor is also MCP-ready, so AI assistants can query Trends through it. One item per term keeps downstream mapping trivial.