Google Trends Scraper

We built Google Trends Scraper to give you Google Trends data as clean, structured results: compare search interest over time and by region, find related searches, and collect what is trending right now, without copying charts by hand.

Instead of managing browsers, cookies, or proxies yourself, choose a mode, enter your search terms, and get one ready-to-use row per result.

Unofficial Actor: This is an independent tool and is not affiliated with, authorized, or endorsed by Google.

---

What you can do with this Actor

1. Explore search interest (explore)

Get the same three result types you see on the Google Trends Explore page, for one search or a batch of up to 100:

• Interest over time: a timeline of search interest from 0 to 100 for each term.
• Interest by region: how interest is spread across countries, subregions, metros, or cities.
• Related queries: the top and rising searches related to each term.

Each line in Search queries is one Google Trends comparison. Put up to 5 terms on one line, separated by commas, to compare them on one scale, just like adding terms on the website. Add more lines to run a batch of independent searches in one run.

You can narrow every search by location, time range (every website preset or your own dates), category, and search type (web, image, news, YouTube, or Google Shopping). Turn off result types you don't need to pay less.

2. Collect Trending Now searches (trendingNow)

Get the searches trending in a country over the past 4 hours, 24 hours, 48 hours, or 7 days, in Google's order. Each trend includes its search volume, increase, start and end time, categories, related searches, and up to 30 news articles.

3. Look up categories and locations (exploreCategories and exploreGeos)

Get the complete lists of Google Trends categories and locations. Use them to find the category ID or location code for your Explore and Trending Now runs. These modes need no other input.

---

Quick Start

1. Click Try for free on the Actor page.
2. Leave Mode on Explore search interest. The input is prefilled with a comparison of bitcoin and ethereum in the United States.
3. Click Start.
4. Open the Dataset tab when the run finishes.

For your first run, you can use the prefilled input:

{
  "mode": "explore",
  "geo": "US",
  "searchQueries": ["bitcoin, ethereum"]
}

This returns one interest-over-time row and one interest-by-region row for the comparison, plus one related-queries row for each term: four rows in total.

---

Input Guide

Only the fields used by your selected mode are read; the others are ignored. The Actor checks your input before it starts. If something is invalid, the run fails right away with a message explaining what to fix, and you are not charged.

Location (geo)

Location is used by both explore and trendingNow and defaults to US. Enter a Google Trends location code:

• A country, such as US, GB, or ID.
• A subregion, such as US-CA.
• A metro area, such as US-CA-807.
• WW for worldwide. Worldwide is available in explore only, because Google Trends has no worldwide Trending Now list.

Run List Explore locations (exploreGeos) to see every supported code. Each row's code can be used directly as Location.

Explore inputs

Input	What it does
searchQueries	1–100 lines. Each line is one comparison of 1–5 comma-separated terms.
timeRange	now 1-H, now 4-H, now 1-d, now 7-d, today 1-m, today 3-m, today 12-m (default), today 5-y, all (2004 to present), or custom.
customTimeRange	Required when timeRange is custom: a start and end date as YYYY-MM-DD YYYY-MM-DD.
category	A Google Trends category ID. 0 (default) means all categories. Run List Explore categories to see every ID.
property	The Google search the interest comes from: "" web search (default), images, news, youtube, or froogle (Google Shopping).
includeInterestOverTime, includeInterestByRegion, includeRelatedQueries	Turn each result type on or off. All are on by default; keep at least one on.
regionResolution	The level used for interest by region: "" Google's default, COUNTRY, REGION (subregion), DMA (metro), or CITY.

Location, time range, category, and search type apply to every line in the run. To compare searches with different settings, use separate runs.

This example compares three terms in the United States over a custom date range and returns only interest over time and related queries:

{
  "mode": "explore",
  "geo": "US",
  "searchQueries": ["chatgpt, gemini, claude"],
  "timeRange": "custom",
  "customTimeRange": "2025-01-01 2025-06-30",
  "includeInterestByRegion": false
}

This example runs a batch of three separate searches worldwide over the past 5 years:

{
  "mode": "explore",
  "geo": "WW",
  "searchQueries": ["running shoes", "hiking boots", "yoga mat"],
  "timeRange": "today 5-y"
}

Trending Now inputs

Input	What it does
geo	The country to list trends for. WW is not supported.
hours	4, 24 (default), 48, or 168 (7 days).
maxArticlesPerTrend	News articles to return for each trend, 0–30. Defaults to 3. When hours is 168, at most 3 is allowed.

{
  "mode": "trendingNow",
  "geo": "GB",
  "hours": "24",
  "maxArticlesPerTrend": 5
}

Category and location lists

{
  "mode": "exploreCategories"
}

Use "mode": "exploreGeos" for the location list.

---

Data Structure & Examples

Every row has a type field that tells you what it contains: interestOverTime, interestByRegion, relatedQueries, trend, category, geo, or error. The Dataset tab has a view for each mode, and the Explore views are split by result type. Every view also shows the type and message columns, so you can spot error rows.

Explore rows always include the settings they were made with (terms, geo, timeRange, category, property) and resolvedTime, the actual date range Google used. Each row therefore stands on its own, even in a large batch.

Results keep Google's order and are not deduplicated. Explore rows follow the order of your input lines. Times are in UTC, and labels such as formattedTime are in English.

JSON exports keep the nested structure shown below. Tabular exports such as CSV or Excel flatten or serialize nested fields.

Interest over time (interestOverTime)

This is an abridged preview, not a complete Dataset row.

{
  "type": "interestOverTime",
  "terms": ["bitcoin", "ethereum"],
  "geo": "US",
  "timeRange": "today 3-m",
  "resolvedTime": "2026-07-14 2026-10-14",
  "timeline": [
    {
      "timestamp": "2026-07-14T00:00:00Z",
      "formattedTime": "Jul 14, 2026",
      "values": [66, 11],
      "formattedValues": ["66", "11"],
      "hasData": [true, true],
      "partial": false
    }
  ],
  "averages": [53, 8]
}

• values holds one number per term, in the order of terms.
• partial is true for the latest point when Google's data for that period is still incomplete.
• averages is filled only for comparisons of two or more terms.

Interest by region (interestByRegion)

This is an abridged preview, not a complete Dataset row.

{
  "type": "interestByRegion",
  "terms": ["bitcoin"],
  "geo": "US",
  "resolution": "REGION",
  "regions": [
    {
      "geoName": "Kansas",
      "geoCode": "US-KS",
      "values": [100],
      "formattedValues": ["100"],
      "hasData": [true],
      "maxValueIndex": 0
    }
  ]
}

• resolution is the level Google used.
• Country, subregion, and metro regions have a geoCode. City regions have coordinates (lat, lng) instead.

Related queries (relatedQueries)

There is one row for each term, so a comparison of three terms returns three related-queries rows.

This is an abridged preview, not a complete Dataset row.

{
  "type": "relatedQueries",
  "terms": ["bitcoin"],
  "term": "bitcoin",
  "top": [
    {
      "query": "bitcoin price",
      "value": 100,
      "formattedValue": "100",
      "exploreUrl": "https://trends.google.com/trends/explore?q=bitcoin+price&date=today+3-m&geo=US"
    }
  ],
  "rising": [
    {
      "query": "british man recovers lost bitcoin",
      "value": 12250,
      "formattedValue": "Breakout",
      "exploreUrl": "https://trends.google.com/trends/explore?q=british+man+recovers+lost+bitcoin&date=today+3-m&geo=US"
    }
  ]
}

• top values are relative interest from 0 to 100.
• rising values show the increase. Google shows them as a percentage such as +1,700%, or as Breakout for very large increases.

Trend (trend)

This is an abridged preview, not a complete Dataset row.

{
  "type": "trend",
  "title": "joe buck",
  "exploreTerm": "joe buck",
  "geo": "US",
  "startedAt": "2026-10-14T00:00:00Z",
  "endedAt": null,
  "active": true,
  "searchVolume": 20000,
  "increasePercentage": 1000,
  "categories": [{ "id": 4, "name": "Entertainment" }],
  "breakdown": ["joe buck", "who is joe buck"],
  "articles": [
    {
      "title": "Michelle Beisner-Buck joined MNF long before Joe Buck made it a family affair",
      "url": "https://www.themirror.com/sport/american-football/michelle-joe-buck-mnf-family-2045760",
      "source": "The Mirror US",
      "publishedAt": "2026-10-05T11:00:00Z",
      "imageUrl": "https://encrypted-tbn2.gstatic.com/images?q=tbn:..."
    }
  ]
}

• searchVolume is a lower bound: 20000 is what the website shows as "20K+".
• active is true while the trend is still trending; endedAt is then null.
• exploreTerm is the form of the trend you can use as an Explore search term.
• breakdown lists the related searches Google groups under the trend.
• Trending Now has no category or status filter. Use categories and active to filter the results yourself.

Category (category) and location (geo)

{
  "type": "category",
  "id": 184,
  "name": "Celebrities & Entertainment News",
  "parentId": 3,
  "path": ["Arts & Entertainment", "Celebrities & Entertainment News"],
  "depth": 2
}

{
  "type": "geo",
  "code": "US-AL-630",
  "id": "630",
  "name": "Birmingham AL",
  "parentCode": "US-AL",
  "path": ["United States", "Alabama", "Birmingham AL"],
  "depth": 3
}

Rows follow Google's order, with each parent before its children. The lists start with "All categories" (0) and "Worldwide" (WW). A category that appears under more than one parent is listed once for each, with a different path.

Error rows (error)

When a search or result type fails, the Actor writes an error row with the settings involved and the error message:

{
  "type": "error",
  "message": "fetch relatedQueries widget RELATED_QUERIES_1: unexpected HTTP status 400",
  "widget": "relatedQueries",
  "term": "ethereum",
  "terms": ["bitcoin", "ethereum"],
  "geo": "WW",
  "timeRange": "today 12-m",
  "category": 0,
  "property": ""
}

In explore, errors are handled per line:

• The search fails: the line gets one error row without a widget field, and the run continues with the next line.
• A result type fails: that result type gets an error row with its widget (and term for related queries), and the line's other result types are still saved.

The run fails only if the search fails for every line. A line whose search succeeds but whose result types all fail returns only error rows, and the run still succeeds. A successful run can therefore contain no results, so check for error rows rather than relying on the run status alone.

In the other modes, a failure writes one error row and fails the run. Error rows are never charged.

---

Operational & Billing Details

Managed Proxies

Requests on Apify are automatically routed through managed residential proxies. You don't need any proxy setup, credentials, or a Google account. This helps keep runs from being blocked, but it can't guarantee that Google Trends is always available.

Pay-per-Event Billing

This Actor uses Apify's Pay Per Event (PPE) model. You pay for the results you receive, not for compute time:

• explore-result: charged once for each interest-over-time, interest-by-region, or related-queries row.
• trend-result: charged once for each Trending Now trend row.
• reference-list: charged once for each successful category or location list run, regardless of how many rows it returns.

For example, one search term with all three result types returns three explore-result rows. A comparison of three terms returns five: one over time, one by region, and three related queries.

What is and isn't charged:

• Rows without data are charged. A search with too little interest still returns its rows with empty arrays, and they count as results.
• Error rows and invalid input are never charged.
• Turn off result types you don't need to pay only for the rows you use.
• Maximum cost per run: If you set one, the Actor saves only the results that fit within it, then finishes successfully and logs that the maximum was reached.

Check the Actor's Pricing tab for current rates.

---

Use with Apify MCP

You can run this Actor from AI assistants through Apify's hosted Model Context Protocol (MCP) server and compatible clients (such as Claude Desktop):

https://mcp.apify.com?tools=crawlstone/google-trends-scraper

Once connected, you can ask your assistant to compare searches, check regional interest, or list what is trending.

---

Frequently Asked Questions

Why do values differ between runs or from the website?

Google Trends values come from a sample of Google searches, and Google can draw a different sample each time you ask, on the website too. The Actor returns the same data the website shows for your request, so two runs with the same input can return slightly different numbers.

Can I compare values across different lines?

No. Google scales each line from 0 to 100 on its own, so a value of 50 in one line doesn't mean the same as 50 in another. To compare terms on one scale, put them in the same line, up to 5 terms.

Why did a search return empty data or an error row?

Searches with too little interest return their rows with empty arrays, as the website shows "not enough data". These rows are still charged. Some combinations are not supported by Google, such as a region resolution that doesn't fit the location, or an unknown category or location code. They return an uncharged error row. In explore, the run continues with your other lines, and it succeeds even if no line returns data, as long as at least one line's search went through. See Error rows under Data Structure & Examples.

Are duplicates removed?

No. Results keep Google's order, and duplicates stay as they are. If you enter the same line twice, you get its results twice.

Why is the article count limited for the past 7 days?

Over the past 7 days, Google's Trending Now lists are large enough that requesting many articles for every trend fails. The Actor therefore allows at most 3 articles per trend when Trending period is Past 7 days.

---

Support

If you need help, open a ticket on this Actor's Discussion tab and include:

• Your Apify Run ID.
• The selected Mode and the Input you used.
• What you expected and what happened instead.

Do not share your Apify API token or other private credentials in support tickets.