Free Google Trends API Scraper — Interest Over Time, Related Queries, Regional Breakdown

A free Google Trends API alternative that pulls keyword interest over time, related queries, related topics, and regional interest for any search term in any country. Hosted Apify Actor — no captchas to solve, no IPs to rotate, no pytrends rate limits to debug. Pay only for the trend queries you run.

What you get

For every keyword you submit:

• Interest over time — full timeseries, 0-100 normalized, with is_partial flag so you can drop the unfinished current period
• Related queries (top + rising) — the queries Google associates with your keyword, ranked by search interest plus a rising list with extracted_value percentages
• Related topics (top + rising) — entity-level associations (Knowledge Graph topics), useful for topical clustering and content strategy
• Regional breakdown — interest by country/region (geo_code, geo_name, value, formatted_value, has_data) when you toggle the flag on
• Configurable timeframe — pytrends-format strings (today 12-m, today 5-y, now 7-d) or explicit YYYY-MM-DD YYYY-MM-DD ranges
• 50+ countries and languages — pass geo=NL, hl=nl-NL, etc.; leave geo blank for worldwide
• Built-in retry-on-429 — rotates a fresh Google session from the cookie pool on every rate-limited widget call (up to 8 retries)
• Optional Apify residential proxy — opt-in for the related-topics widget which is consistently rate-limited from datacenter IPs

Why scrape Google Trends instead of using pytrends

The obvious free alternative is pytrends, the unofficial Python library that's been the de-facto Google Trends client since 2016. It works — until it doesn't.

Rate limits. Google's /trends/api/widgetdata/* endpoints throttle per (cookie, IP) pair. A bare pytrends client starts eating 429s within a few dozen requests. The fix is maintaining a pool of fresh Google sessions, rotating proxies, and implementing backoff — and then updating that infrastructure every time Google tweaks the protocol. A side project, not a weekend script.

The related-topics widget specifically. RELATED_TOPICS is the strictest of all the Trends widgets and most pytrends installs get empty data on the first call. This Actor handles it: synthetic GA/utm cookies, Chrome 131 TLS impersonation via curl_cffi, optional Apify residential proxy for the hardest queries.

The Apify cloud also frees you from running a long-lived scraper on your laptop. Submit 100 keywords, walk away, come back to a clean JSON dataset. No Heroku worker, no cron tab, no captcha-solving service.

Input

Field	Default	Description
keywords	required	Array of search terms — one trend query per keyword
geo	US	Two-letter country code (US, GB, NL, DE); empty = worldwide
hl	en-US	Language code (en-US, nl-NL, de-DE)
timeframe	today 12-m	pytrends format: today 12-m, today 5-y, now 7-d, or YYYY-MM-DD YYYY-MM-DD
include_related_queries	true	Top + rising related search terms
include_related_topics	true	Top + rising related topics (entities)
include_regional_interest	false	Country/region breakdown (slower)
use_cookies	true	Rotate a fresh session per widget call
use_proxy	false	Route through the Tim IPv6 proxy (default off — direct usually works better)
use_residential_proxy	false	Use Apify residential proxy for the hardest related-topics calls (incurs Apify proxy CUs)

Output

{
  "keyword": "pizza",
  "geo": "US",
  "hl": "en-US",
  "timeframe": "today 12-m",
  "interest_over_time": [
    {"date": "May 5, 2025", "value": 78, "is_partial": false},
    {"date": "May 12, 2025", "value": 81, "is_partial": false}
  ],
  "related_queries_top": [
    {"query": "pizza near me", "value": 100, "extracted_value": 100,
     "link": "/trends/explore?q=pizza+near+me", "formatted_value": "100"}
  ],
  "related_queries_rising": [
    {"query": "detroit style pizza", "value": 250, "extracted_value": 250,
     "link": "/trends/explore?q=detroit+style+pizza", "formatted_value": "+250%"}
  ],
  "related_topics_top": [...],
  "related_topics_rising": [...],
  "interest_by_region": [
    {"geo_code": "US-NY", "geo_name": "New York", "value": 100,
     "formatted_value": "100", "has_data": true}
  ]
}

Use cases

SEO content strategy — pull rising related queries for your money keywords, sort by extracted_value, and build a content calendar around gaining queries before competitors notice. Cheaper than Ahrefs Trending Topics and updates daily.

Ecommerce demand forecasting — track interest over time for product categories across countries to time inventory buys and plan launch windows. Combine interest_over_time with regional breakdown to spot the country where a product is taking off 12-16 weeks before peak.

Content strategist — feed related_topics_rising into a topical-cluster brief so writers cover the entities Google's Knowledge Graph associates with the seed keyword. The 5-year timeframe doubles as a seasonality detector.

Marketing analyst / agency — run nightly trend tracking across a 50-200 keyword client portfolio and surface anomalies. The cookie pool

• residential-proxy fallback makes batch jobs pytrends can't sustain actually feasible.

How it compares

Tool	Pricing model	Maintenance	Hosted?
pytrends (Python lib)	Free	You debug rate limits, captchas, cookie expiry	No — runs on your box
Google Trends API Alpha	Closed alpha, application-gated	Google manages	Yes
DataForSEO Trends API	$0.0006-$0.002/req minimums + plan	Vendor handles	Yes
Bright Data SERP / Trends	$1.50-$3.50/1k requests + proxy	Vendor handles	Yes
SerpApi Google Trends	$50-$250/mo plans	Vendor handles	Yes
This Actor	$0.005/trend_query, no plan	Apify handles	Yes

DataForSEO, Bright Data, and SerpApi all work — they're just on subscription plans that don't suit ad-hoc use. This Actor is pay-per-event on Apify: pay for the queries you run, walk away when you stop.

Pricing

$0.005 per trend_query (one keyword = one event). Includes interest over time, related queries, related topics, and optional regional breakdown in the same charge.

No actor-start fee. No compute units billed separately. The optional residential proxy adds Apify proxy compute units at $1.20/GB only when you explicitly toggle use_residential_proxy=true.

Limits and gotchas

• Timeframe normalization — values are normalized 0-100 within the timeframe. Two timeframes for the same keyword aren't directly comparable; use a fixed window when benchmarking
• Sparse keywords return zeros — under ~50 monthly searches in your geo, Trends returns all-zero timeseries (Google policy, not a scraper bug)
• Related topics can be flaky — toggle use_residential_proxy=true for highest reliability on the topics widget specifically
• Regional breakdown adds latency — each country pull is a separate widget call; expect ~3-5x runtime when enabled
• is_partial: true rows — the most recent period is provisional; drop or annotate for analysis
• One keyword per request — Google's /explore accepts up to 5 keywords per comparisonItem but the values rescale; we send one keyword per request for consistent cross-keyword scaling
• No sub-7-day granularity beyond now ranges — Google returns hourly data only for now timeframes; older ranges get weekly/monthly buckets

FAQ

Does this use the official Google Trends API Alpha? No — that API is in closed alpha and gated behind an application. This Actor scrapes the same /trends/api/widgetdata/* endpoints that the trends.google.com web UI uses, with Chrome 131 TLS impersonation and a rotating cookie pool to avoid rate limits.

Can I get hourly trend data? Yes, for the now 7-d, now 4-H, and now 1-H timeframes. Anything older than 7 days returns weekly buckets (timeframes up to 5 years) or monthly buckets (all and longer ranges) — that's a Google constraint, not a scraper limitation.

How do I compare interest across multiple countries? Run the same keyword with different geo codes in separate calls. Each result is normalized 0-100 within its own geo, so for true cross-country comparison either (a) use the regional breakdown of one global call, or (b) check the extracted_value fields rather than the raw value.

Why does the related-topics widget sometimes return empty? The topics widget has the strictest rate limits on Trends. We retry up to 8 times with fresh cookies, but if your keyword has very low volume or Trends is heavily throttling that day, the array will be empty. Toggle use_residential_proxy=true for the highest success rate.

Is the data the same as what trends.google.com shows? Yes — same widget endpoints, same req payloads. The only difference is that this scraper bundles all four widgets (timeseries, related queries, related topics, regional) into a single result object per keyword.

Related Actors

• Free Keyword Research Tool — monthly search volume, CPC, SEO difficulty, and intent for any seed keyword across 50+ countries. Pair with this Actor: use Trends to spot rising queries, then keyword-research them for volume + difficulty.
• Free Google News API — pull news articles for any keyword. Pair with Trends to correlate news cycles with search interest spikes.
• Free Google AI Overview Scraper — track which queries surface AI Overviews and which domains Google's AI cites. Combine with Trends to find rising queries where AI Overview cannibalisation is most acute.