Google Trends Explore

Google Trends Explore is a Google Trends Explore API that turns a classic Explore comparison into up to five typed, 18-field Dataset rows with timelines, geographic interest, and top and rising related queries.

• One Actor call, five terms, one normalized scale — all terms share the same comparison, so their 0-100 values are directly comparable with each other.
• The response states its own shape: the absolute window Google resolved the request to, and the bucket size it chose, in every record.
• Geography switches the breakdown: Worldwide returns country-level rows, a single country returns that country's subregions.
• Every related query carries its own Explore link, so a term worth investigating is one request away from its own comparison.

Run a one-result test · View API

The smallest useful run uses one term and bills one Result event plus one Actor Start event — $0.02350 on the FREE tier. A one-term row leaves interest_average null; add a second term when the average comparison matters.

What does Google Trends Explore do?

Google Trends Explore accepts terms, geography, window, category, and search surface, then returns the source comparison as typed records. Each distinct source keyword can become one stable 18-field Dataset row; source-empty terms are skipped, while a row with one measured panel can keep another unavailable panel as null.

It does not return search volumes: Google's Trends data documentation defines the 0-100 values as relative interest, not counts. Separate requests are normalized independently. This Actor is not Google's access-gated Trends API and has no affiliation with Google; it reads the public classic Explore experience.

How do I run Google Trends Explore?

Fill the five inputs, start the run, and read one row per term from the default Dataset:

1. Enter one to five terms in keywords.
2. Choose location, time_range, category and search_type.
3. Start the run and read the default Dataset.

This two-term scenario exercises the shared comparison scale; delete either term for the smallest one-result test:

{
  "keywords": ["bitcoin", "ethereum"],
  "location": "Worldwide",
  "time_range": "Past 90 days",
  "category": "Finance",
  "search_type": "News search"
}

Billing follows published rows: five distinct source terms can produce five rows and five Result events. A source-empty term is skipped, and two requested values that Google normalizes to the same source keyword are emitted once rather than billed twice.

What data does Google Trends Explore return?

Each published record has 18 fields.

Group	Fields	Notes
Term and scope	keyword, location, category, search_type, platform	Read back from what Google resolved the request to, not echoed from the input.
Resolved scope	date_start, date_end, resolution, geo_resolution, rising_start, rising_end	Timeline grain, geographic grain, and the previous window used for rising-query growth.
Measurements	interest_average, timeline, regions, queries_top, queries_rising	interest_average is Google's own mean and is supplied only for a comparison of two or more terms.
Provenance	processor, processed_at	Which Actor produced the record, and when.

Illustrative abbreviated record from the scenario above:

{
  "keyword": "bitcoin",
  "location": "Worldwide",
  "category": "Finance",
  "search_type": "News search",
  "date_start": "2026-05-09",
  "date_end": "2026-08-07",
  "resolution": "DAY",
  "geo_resolution": "COUNTRY",
  "rising_start": "2026-02-08",
  "rising_end": "2026-05-08",
  "interest_average": 70,
  "timeline": [{ "time": "2026-07-01T00:00:00+00:00", "value": 64 }],
  "regions": [{ "name": "Singapore", "code": "SG", "value": 56 }],
  "queries_top": [{ "query": "bitcoin price", "value": 100, "url": "https://trends.google.com/trends/explore?q=bitcoin+price&date=today+3-m" }],
  "queries_rising": [{ "query": "bitcoin etf news", "value": 13500, "url": "https://trends.google.com/trends/explore?q=bitcoin+etf+news&date=today+3-m", "is_breakout": true }]
}

Values Google does not supply stay empty rather than being filled in: interest_average is null on a single-term run, and an unavailable panel remains null with the incomplete row counted in the run summary. A source no-data timeline bucket keeps its timestamp with value: null; a measured value below one keeps value: 0 and value_text: "<1".

In queries_rising, value is percentage growth — 13500 means +13,500% — and is_breakout marks the terms Google labels "Breakout" instead of printing a number. On multi-year windows a timeline entry may carry a note recording a change Google made to how it measured from that point on; the series is not directly comparable across such a point.

What inputs can I configure?

All five inputs are required.

Input	Type	Required	Effect
keywords	array, 1-5	Yes	The terms to compare. A sixth is refused at the source.
location	select, 236 options	Yes	Worldwide or one listed country or territory. Also decides what regions contains.
time_range	select, 9 options	Yes	Past hour through 2004-present. Determines resolution.
category	select, 26 options	Yes	Applies to the whole request, including regions and related queries.
search_type	select, 5 options	Yes	Web, Image, News, Shopping or YouTube search.

What platforms and markets does Google Trends Explore cover?

One source: Google Trends, read through its public Explore experience.

• Markets — 236 schema options including Worldwide. Worldwide returns country-level rows in regions; a selected country returns its own subregions, so geography selects the breakdown as well as the measurement.
• Surfaces — Web, Image, News, Shopping and YouTube search. Each is an independent index; the same term can rank very differently between them and the figures do not translate.
• Windows — nine, from the past hour to the full 2004-present archive. The window fixes the granularity: a 12-month request returns weekly buckets, a 7-day request hourly ones, the full archive monthly ones.
• Categories — 26, applied to the entire request.

Coverage is source-dependent: low-volume terms can return no data, especially in smaller markets. Broaden the term, geography, or window when a narrow combination is empty.

Why use Google Trends Explore?

Because the comparison arrives normalized and self-describing: up to five terms on one 0-100 scale, in fixed 18-field records that state their window, timeline grain, geographic grain, and rising-query baseline.

Capability	What it gives a data team
Five terms normalized in one comparison	A relative ranking, instead of stitching together separately scaled runs that cannot be compared.
Resolved window and resolution in the record	The series is self-describing, so downstream code does not have to infer bucket size from point count.
Geography-aware regions	Country-level or subregion-level breakdown from the same field, chosen by the input rather than by parsing a different panel.
Related queries with Explore links	Discovery and the next request in one step.
Fixed 18-field record	Parsing remains stable when a source panel is unavailable; the affected field stays null.

The trade-off is inherent to the source: because each request is normalized on its own, this Actor cannot give you a series that merges across requests. If your work depends on that, Google's own Trends API is the right tool and this one is not.

Who is Google Trends Explore for?

Developers and data teams use it when Explore must feed a pipeline rather than remain a browser tab. Examples include scheduled term-set tracking, market comparisons across countries or surfaces, and typed answers to "which term is moving?"

Store interest_average for a fixed comparison set, compare one term across Web and Shopping, or use queries_rising as emerging phrasing to review.

It is the wrong tool if you need absolute search volumes, cross-request comparable scaling, more than five terms in one comparison, or keyword difficulty and cost-per-click data. Those are different products, and this Actor does not approximate them.

How can I use Google Trends Explore through the API or MCP?

The Actor ID is ubXiK9s9p31KTeQrg — also addressable by name as truefetch/google-trends-explore; both forms work in the API, the SDKs, and MCP.

curl -X POST "https://api.apify.com/v2/acts/truefetch~google-trends-explore/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords":["bitcoin","ethereum"],"location":"Worldwide","time_range":"Past 90 days","category":"Finance","search_type":"News search"}'

Through MCP, expose it with the Apify MCP server and call it by Actor ID:

{
  "mcpServers": {
    "apify": {
      "command": "npx",
      "args": ["-y", "@apify/actors-mcp-server", "--actors", "truefetch/google-trends-explore"],
      "env": { "APIFY_TOKEN": "YOUR_TOKEN" }
    }
  }
}

Generated SDK snippets and the OpenAPI definition are on the API page. For long or scheduled workloads, start the run asynchronously, keep the run ID, and page the resulting Dataset.

How much does Google Trends Explore cost?

One published term row costs $0.01350 on the FREE tier, plus a $0.01000 Actor Start per run — the smallest useful one-row test is $0.02350.

Event	Billing unit	FREE	BRONZE	SILVER	GOLD / PLATINUM / DIAMOND
Result	One compared term	$0.01350	$0.01215	$0.01080	$0.01013
Actor Start	One per GB of memory, minimum one	$0.01000	$0.01000	$0.01000	$0.01000

A one-row run bills one Actor Start and one Result event: $0.01000 + $0.01350 = $0.02350 on the FREE tier. Prices can change; the pricing page is authoritative.

How does Google Trends Explore compare with alternatives?

Against the Explore interface. Free, and entirely adequate for one question asked once. It stops being adequate when the same comparison has to run on a schedule, across markets, or feed something downstream.

Against Google's own Trends API. The Trends API alpha returns consistently scaled data that can be joined across requests and compared across dozens of terms — both things this Actor cannot do. Access is application-gated. If merging series across requests is required and you have alpha access, choose it.

Against narrower Trends Actors. Several read Google Trends. This one is built around the comparison rather than the single term: the values arrive scaled against each other, and the record states the window and bucket size it was measured at, so a stored series stays interpretable months later.

Against a keyword research suite. Those give volumes, difficulty and CPC, which this Actor does not. They are complementary, not substitutes.

What are the limits and troubleshooting steps?

Each limit below includes its next step:

• A sixth term is refused. This Actor accepts one to five values. Split larger sets, but do not compare values across separately normalized runs.
• A low-volume term returns nothing. Try a broader term, larger market, or longer window.
• One section is empty. The affected field stays null and the run counts an incomplete row. Retry once; include the run ID if it persists.
• Identical runs differ slightly. Treat small changes on low-interest terms as sampling noise.
• The run returns nothing. Google refused the request window; retry later rather than treating it as an empty trend.

Smallest reproducible input for a bug report is the two-term bitcoin/ethereum JSON from the How-do-I-run section. Attach it, the run ID, and which section was missing to Issues.

Frequently asked questions

Is there a Google Trends interest by region API?

Yes. regions follows location: Worldwide returns countries such as SG, while a country returns subregions such as US-WY. Each entry carries the name, source code, and 0-100 value.

Is there a Google Trends related queries API?

Yes. queries_top and queries_rising return query text, value, and the source Explore link. Top values use a 0-100 scale; rising values are percentage growth.

Why is interest_average null?

Google supplies it only for comparisons of two or more terms. A single-term row leaves it null instead of inserting a locally calculated substitute.

Can I compare numbers from two different runs?

No. Each run is normalized independently, so put every term you want to compare into one run.

What does resolution mean and why does it change?

It is Google's timeline bucket size for the selected window: an hour uses minutes, while the full archive uses months.

What is a note on a timeline entry?

It records a source methodology change from that point onward, so the Actor preserves it with the affected bucket.

Does the category filter affect related queries too?

Yes. It applies to the timeline, regions, and both related-query lists.

Is there a free Google Trends Explore API?

No open official API currently matches this workflow. Google's alpha is application-gated; this Actor reads classic Explore and charges $0.02350 for one Actor Start plus one published row on the FREE tier.

Can I schedule runs to track a term set?

Yes. Schedule the same term set, geography, window, and surface. Store interest_average per term, and watch queries_rising for is_breakout entries.

Related TrueFetch Actors

Three siblings, each for a different stage of the same investigation:

• Price Tracker API — for scheduled product-price snapshots when the subject is retail monitoring rather than search-interest research.
• Google Maps Business Listing — turns a market that indexes strongly for a term into the businesses operating there.
• Reddit Search API — checks whether a search-interest signal is also appearing in public Reddit search results.

Support

• Store page
• API reference
• Input configuration
• Pricing
• Report an issue
• TrueFetch community

Run a one-result test · View API

Last Updated: August 12, 2026