Google Trends Scraper Fixed

Why use this Actor?

Google Trends normally leaves you to interpret several disconnected charts by eye: one for interest over time, one for related queries, one for regional breakdown, per keyword, one at a time.

This Actor retrieves the same underlying data for every keyword you give it, and adds one thing the raw Trends UI doesn't: a single opportunity score that combines recent interest, growth rate, acceleration, and volatility into one comparable number. You still get the full underlying time series, related queries, rising queries, and regional data — the score is a summary layer on top, not a replacement for it.

The score is an internal analytical signal, not a guarantee of business demand, revenue, or search volume.

What you get

📈 Interest over time The complete normalized time series for each keyword, at the resolution Google Trends returns for your selected time range.	🚀 Related & rising queries The top related queries and the fastest-growing ("rising") related queries, including breakout terms.
🌍 Regional interest Relative interest by region (state/country level, depending on geo), sorted highest first.	🎯 Opportunity score One 0–100 score per keyword, combining trend level, growth, acceleration, and volatility, so you can rank keywords instead of eyeballing charts.

From raw trends to opportunity intelligence

Raw Google Trends workflow	This Actor
Manually inspect charts	Receive structured time-series data
Compare keywords visually	Rank keywords programmatically
Interpret growth manually	Receive a calculated growth rate
Guess whether momentum is increasing	Receive an acceleration label
Ignore unstable spikes	Apply a volatility penalty
Copy related queries by hand	Receive top and rising queries as arrays
Review one region at a time	Receive structured regional interest

This doesn't replace judgment or predict outcomes — it removes the manual charting step so you can screen more keywords faster.

How it works

Keywords
   |
Google Trends "explore" request
   |
   |-- Interest over time
   |-- Related queries (top + rising)
   +-- Regional interest
   |
Growth + acceleration + volatility scoring
   |
Opportunity score
   |
Structured Apify dataset (one row per keyword + one run summary)

Each keyword is fetched independently through Google Trends' own (unofficial) explore + widgetdata interfaces — the same two-step mechanism the open-source pytrends library uses. There is no official Google Trends API; this Actor calls the same endpoints the public Trends website itself uses.

Opportunity score

opportunityScore
= (trendScore * 0.5)
+ (max(growthRatePct, 0) * 0.3)
- (volatility * 0.2)

Clamped to a 0–100 range. These are the actual weights used in the scoring code, not illustrative figures.

Recent trend level (trendScore)

The average of the most recent 4 data points in the time series, on Google Trends' own 0–100 relative scale.

Growth rate (growthRatePct)

Percentage change between the average of the most recent 4 points and the average of the earliest 4 points in the selected time range.

Acceleration

Compares first-half momentum to second-half momentum within the series (not just overall direction) and labels the result accelerating, steady, or decelerating.

Volatility

The population standard deviation of the interest values. Higher volatility subtracts from the opportunity score, so a keyword that's both elevated and rising, without wild week-to-week swings, scores highest.

The opportunity score is a comparative analytical signal, not an absolute search-volume estimate, revenue forecast, or guarantee of future demand. Scoring requires at least 4 data points in the time series — very short time ranges may not have enough points to score.

Use cases

• Compare several API or product ideas before committing engineering time
• Discover content topics with rising search interest
• Monitor interest in a product category over a season
• Identify which regions show the strongest relative demand for a term
• Compare momentum between a brand name and its competitors
• Surface rising related searches worth targeting
• Prioritize SEO or content topics by growth rate rather than gut feel
• Track seasonal interest patterns year over year
• Build a recurring market-intelligence workflow on a schedule
• Feed structured trend data into your own dashboards, databases, or agents

Input

Field	Type	Required	Default	Description
keywords	array of strings	Yes	—	One or more search terms to analyze.
geo	string	No	Worldwide (blank)	Two-letter country code, e.g. US, GB, DE.
timeRange	enum	No	12m	1h, 4h, 1d, 7d, 1m, 3m, 12m, or 5y.
category	integer	No	0	Google Trends category ID to narrow results; 0 = all categories.
proxyConfiguration	object	No	Apify Proxy enabled	Proxy routing for Google Trends requests (see Proxy configuration).

Field names and defaults match the Actor's live input schema exactly.

Input examples

Compare a few keyword ideas

{
  "keywords": ["AI agents", "MCP servers", "browser automation"],
  "geo": "US",
  "timeRange": "12m",
  "category": 0,
  "proxyConfiguration": { "useApifyProxy": true }
}

One short-term trend, single keyword

{
  "keywords": ["Google Gemini"],
  "geo": "US",
  "timeRange": "7d",
  "category": 0,
  "proxyConfiguration": { "useApifyProxy": true }
}

Worldwide, long-term comparison

{
  "keywords": ["electric vehicles", "hybrid vehicles"],
  "geo": "",
  "timeRange": "5y",
  "category": 0,
  "proxyConfiguration": { "useApifyProxy": true }
}

Output

One dataset row per requested keyword, plus one final run_summary row. Each keyword row is one of three states: COMPLETED (scored normally), NO_DATA (Trends returned no time series for that keyword/geo/timeframe combination), or EXPLORE_FAILED (the initial request itself failed, typically a block or timeout).

Field	Type	Description
keyword	string	The keyword this row is for.
geo	string	Region requested ("worldwide" if left blank).
timeRange	string	Time range requested.
interestOverTime	array	{date, timestamp, value} per data point.
relatedQueriesTop	array	{query, value, formattedValue}, top related queries.
relatedQueriesRising	array	{query, value, formattedValue}, fastest-growing related queries (may include "Breakout").
regionalInterest	array	{geoName, geoCode, value}, sorted highest first.
trendScore	number or null	See Opportunity score.
growthRatePct	number or null	Percentage growth, recent vs. earlier average.
acceleration	string or null	accelerating, steady, or decelerating.
volatility	number or null	Population standard deviation of interest values.
opportunityScore	number or null	0–100 combined score.
status	string	COMPLETED, NO_DATA, or EXPLORE_FAILED.
scrapedAt	string	ISO timestamp.

Output example

Real result from a live run (trimmed to a few data points for readability):

{
  "keyword": "electric vehicles",
  "geo": "US",
  "timeRange": "3m",
  "interestOverTime": [
    { "date": "May 10 – 16, 2026", "timestamp": "1778371200", "value": 62 },
    { "date": "May 17 – 23, 2026", "timestamp": "1778976000", "value": 76 },
    { "date": "May 24 – 30, 2026", "timestamp": "1779580800", "value": 82 },
    { "date": "May 31 – Jun 6, 2026", "timestamp": "1780185600", "value": 85 }
  ],
  "relatedQueriesTop": [
    { "query": "electric car", "value": 100, "formattedValue": "100" },
    { "query": "hybrid vehicles", "value": 72, "formattedValue": "72" }
  ],
  "relatedQueriesRising": [
    { "query": "ev tax credit 2026", "value": 850, "formattedValue": "+850%" }
  ],
  "regionalInterest": [
    { "geoName": "California", "geoCode": "US-CA", "value": 100 },
    { "geoName": "Washington", "geoCode": "US-WA", "value": 61 }
  ],
  "trendScore": 76.3,
  "growthRatePct": 14.2,
  "acceleration": "accelerating",
  "volatility": 9.1,
  "opportunityScore": 14.9,
  "status": "COMPLETED",
  "scrapedAt": "2026-08-01T05:57:00.000000+00:00"
}

Field values here are drawn from a real run against electric vehicles; the interestOverTime array is trimmed for length (a full run returns the complete series for the selected time range).

run_summary row (the actual fields returned, once per run):

{
  "recordType": "run_summary",
  "totalKeywords": 2,
  "completed": 2,
  "failed": 0,
  "topOpportunity": "electric vehicles",
  "scrapedAt": "2026-08-01T05:57:07.506507+00:00"
}

Understanding the results

Interest values follow Google Trends' own scale: relative to the keyword's peak within the requested time range and region, 0–100. They are not absolute monthly search volume. A growthRatePct on a rising-but-small keyword can look dramatic (breakout related queries in particular can show huge percentage growth off a tiny base) — read percentage growth alongside the absolute trendScore, not in isolation.

Status values

Status	Meaning	Recommended action
COMPLETED	Trend data was retrieved and scored.	Use the result normally.
NO_DATA	Google Trends returned no usable time series for this keyword/geo/timeframe combination.	Try a different keyword, region, category, or time range.
EXPLORE_FAILED	The initial Trends request itself failed.	Retry, ideally with residential proxy routing (see below).

Proxy configuration

Google Trends frequently rate-limits or blocks requests from datacenter/cloud IP ranges — this affects any tool that calls it, including this Actor and other Trends scrapers on the Store. Apify Proxy is enabled by default (useApifyProxy: true); routing through the residential proxy group can improve reliability further:

{
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Proxy use reduces blocking but can't guarantee every keyword succeeds on every run. Keywords that come back EXPLORE_FAILED or NO_DATA are not charged (see Pricing) and can simply be re-submitted.

Pricing

Pay-per-event pricing: one Actor start event per run, plus one keyword-processed event for each keyword that reaches COMPLETED status. Keywords that end in NO_DATA or EXPLORE_FAILED are not charged.

5 submitted keywords
4 completed successfully
1 failed (EXPLORE_FAILED)

Billable events:
- 1 Actor start
- 4 keyword-processed events

Current per-event prices are shown in the Pricing tab on this Actor's Store page.

Notes and limitations

• Google Trends has no official API and no published SLA for this unofficial interface — occasional blocks or failures are expected, not a sign of a broken Actor.
• Results vary by region, time range, category, and exact query wording; two close phrasings of the same idea can score differently.
• Values use Google Trends' relative 0–100 scale, not absolute search volume.
• Scoring requires at least 4 time-series data points; very short time ranges may return fewer and skip scoring (trendScore, growthRatePct, acceleration, volatility, and opportunityScore will all be null).
• Short time ranges are inherently more volatile than long ones.
• Rising/breakout related queries can show very large percentage growth from a small historical base — treat them as directional signals, not precise multipliers.
• Some keyword/region/category combinations return no related-query or regional data even when the main time series succeeds.
• A keyword that fails on one run can succeed on a re-run, particularly with a different proxy route.

Recommended workflow

Run a batch of candidate keywords together, sort the results by opportunityScore, then look at relatedQueriesRising and regionalInterest on your top candidates before deciding where to invest further research.

Frequently asked questions

Is this an official Google API? No. It uses Google Trends' unofficial explore and widgetdata interfaces — the same ones the public Trends website itself calls.

Is the opportunity score a search-volume estimate? No. It's a comparative analytical score derived from relative trend behavior (level, growth, acceleration, volatility), not a volume or revenue figure.

Why did a keyword return NO_DATA? The term likely has insufficient recorded search activity for the specific region, category, and time range combination you selected.

Why did a keyword return EXPLORE_FAILED? The request was likely blocked or rate-limited. Retry, ideally with the residential proxy group enabled.

Can I compare unrelated keywords against each other? Yes, but interpret carefully — Trends values are relative to each keyword's own peak, so a score comparison across very different topics needs business context, not just the raw numbers.

Are failed keywords charged? No — only keywords that reach COMPLETED status trigger the keyword-processed charge.

Can I run this on a schedule? Yes, using Apify's standard Schedules feature (available to any Actor on the platform) — nothing Trends-specific is required to set that up.

Can agents call this Actor? Yes, through the standard Apify API/SDK like any Actor, and through Apify's general-purpose MCP server (which can call any Actor on the platform by ID). This Actor does not have its own dedicated MCP/Standby endpoint.

Run via API

curl -X POST \
  "https://api.apify.com/v2/acts/ayeeyee~google-trends-scraper-fixed/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "keywords": ["AI agents", "MCP servers"],
    "geo": "US",
    "timeRange": "12m",
    "category": 0,
    "proxyConfiguration": { "useApifyProxy": true }
  }'

Related Actors

Part of a broader catalog of rebuilt-for-real actors on this account, including LinkedIn Jobs Scraper Enhanced and a five-Actor security-scanning line (Gitleaks, Trivy, Syft, Grype, OSV-Scanner).

Start comparing opportunities

Add the keywords you're considering, pick a market and time range, and run the Actor. The resulting dataset gives you both the underlying Google Trends data and a consistent opportunity score for deciding what to investigate next.