Google Trends Scraper & API: interest over time, regions, related queries, Trending Now

Get Google Trends data as clean JSON, CSV or Excel. Compare up to 5 keywords on the same 0-100 scale, for any country or region, any time range, any category and any Google search type: web, news, images, YouTube or Shopping. For each search term you get:

• Interest over time, with average, peak and latest values worked out for you
• Interest by region: countries, states, metros (DMA) or cities, plus each term's share per region when you compare terms
• Related queries, top and rising, with Breakout flags
• Related topics, top and rising, when Google returns them (see the FAQ)
• Trending Now: today's trending searches by country, with approximate search volume and the news stories behind them

$3 per 1,000 search terms and $1 per 1,000 trending searches. There is no start fee and no monthly rental. Terms that fail or have no data are free.

Docs, examples, MCP setup and an honest comparison with other actors: plugin.benjaminbritton.co.uk/google-trends-api/

Why use this Google Trends API

• Maintained, with a clear status. The actor follows the same request flow as the Trends website: explore, then widget tokens, then widget data. It paces its requests and backs off and retries when Google rate-limits. When Google changes something, the fix ships here.
• You pay for results, not attempts. A comparison that fails after all retries, or a term Google has "not enough data" for, is not charged. RUN_SUMMARY lists exactly what happened to every comparison.
• Real comparisons. Put chatgpt, gemini, claude on one line and all three come back on one shared scale, the way the Trends page shows them. Each term gets its own row, so you can sort and filter the results.
• Two output shapes, same price. Choose one row per search term (nested, best for APIs and AI agents) or one row per data point (long format that opens directly in Excel, Google Sheets, Looker Studio or Power BI). You are charged per search term either way, never per row.
• Every Trends option. Choose the geo (country or region), time range (past hour to 2004-present, or a custom date range), category, search type, region resolution, low-volume regions and language.
• Many countries in one run. Set geo to US, GB, DE and every comparison runs for each country.
• No personal data. Google Trends data is aggregated and anonymised. The actor doesn't log in and collects nothing about individuals.

Use cases

• SEO and content planning: find rising queries and breakout topics before they peak, and check seasonality before you commit to a content calendar.
• Brand and competitor tracking: compare your brand with 4 competitors every week, per country.
• Product and market research: check demand for a product or category by region before launch, and compare demand across YouTube, Shopping and web search.
• Newsrooms and social teams: pull Trending Now for several countries every morning, with the news links behind each trend.
• AI agents and automations: call it from Make, Zapier, n8n, LangChain or Apify's MCP server, and get compact JSON with summary fields (averageInterest, peakDate, latestInterest) that an LLM can reason over.
• Data science: build long-run interest time series as features for forecasting models.

How to use it

1. Add Search terms, one comparison per line. Put up to 5 comma-separated terms on a line to compare them, or a single term to fetch it on its own. Google topic IDs (such as /m/0dl567) work too.
2. Set Country / region (leave it empty for Worldwide), the Time range, and optionally the Category and Search type.
3. Optionally add Trending Now countries.
4. Click Start, then download the dataset as JSON, CSV, Excel or XML, or read it with the Apify API.

Input example: compare three AI assistants in the US over 12 months

{
  "searchTerms": ["chatgpt, gemini, claude"],
  "geo": "US",
  "timeframe": "today 12-m"
}

Input example: YouTube interest in two phones, in the UK and Germany, as spreadsheet rows

{
  "searchTerms": ["iphone 17, pixel 10"],
  "geo": "GB, DE",
  "timeframe": "today 3-m",
  "property": "youtube",
  "outputFormat": "rows"
}

Input example: custom date range, Computers & Electronics category, metro-level regions

{
  "searchTerms": ["python", "rust, golang"],
  "geo": "US",
  "customTimeframe": "2024-01-01 2024-12-31",
  "category": 5,
  "regionResolution": "DMA"
}

Input example: Trending Now for five countries (no search terms needed)

{
  "trendingGeos": ["US", "GB", "DE", "IN", "JP"]
}

Input fields

Field	What it does
searchTerms	One comparison per line. Up to 5 comma-separated terms per line share one 0-100 scale. A line with more than 5 terms is split into groups of 5, and you get a warning
geo	US, GB, US-CA, GB-ENG... Empty means Worldwide. Separate several with commas to run each comparison per geo
timeframe	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
customTimeframe	Overrides timeframe: 2024-01-01 2024-12-31, or hourly 2026-11-01T00 2026-11-07T23
category	Trends category ID. 0 means all categories. Examples: 5 Computers & Electronics, 7 Finance, 18 Shopping, 20 Sports, 45 Health, 71 Food & Drink, 958 Jobs & Education. Sub-category IDs work too, for example 31 Programming
property	"" web (default), news, images, youtube, froogle (Google Shopping)
includeInterestOverTime, includeInterestByRegion, includeRelatedQueries, includeRelatedTopics	Turn sections off for faster runs. The price per term stays the same
regionResolution	auto, COUNTRY, REGION, DMA (US metros), CITY
includeLowSearchVolumeRegions	Also list low-volume regions
trendingGeos	Country codes for Trending Now
outputFormat	perTerm (default) or rows (long format)
language	en-US (default), de, fr, ja... Sets the language of region names and topic titles
proxyConfiguration	Apify Proxy is on by default. If Google rate-limits a request, the actor retries it on a new proxy IP
requestDelaySecs, maxRetries	Request pacing and retry budget

Output

One row per search term (default)

{
  "type": "searchTerm",
  "searchTerm": "chatgpt",
  "comparedWith": ["gemini", "claude"],
  "termIndex": 0,
  "geo": "US",
  "timeframe": "today 12-m",
  "category": 0,
  "property": "web",
  "hasData": true,
  "averageInterest": 74.0,
  "peakInterest": 100,
  "peakDate": "2025-11-02",
  "latestInterest": 76,
  "latestDate": "2026-11-01",
  "interestOverTime": [
    {"time": "2025-11-02T00:00:00Z", "date": "2025-11-02", "formattedTime": "Nov 2 – 8, 2025", "value": 100, "hasData": true, "isPartial": false}
  ],
  "interestByRegion": [
    {"geoCode": "US-CA", "geoName": "California", "value": 100},
    {"geoCode": "US-WY", "geoName": "Wyoming", "value": 89}
  ],
  "interestByRegionShare": [
    {"geoCode": "US-CA", "geoName": "California", "share": 57}
  ],
  "relatedQueriesTop": [
    {"rank": 1, "query": "chatgpt ai", "value": 100, "formattedValue": "100", "exploreUrl": "https://trends.google.com/trends/explore?q=chatgpt+ai&date=today+12-m&geo=US"}
  ],
  "relatedQueriesRising": [
    {"rank": 1, "query": "chatgpt codex", "value": 1300, "formattedValue": "+1,300%", "isBreakout": false, "exploreUrl": "https://trends.google.com/trends/explore?q=chatgpt+codex&date=today+12-m&geo=US"}
  ],
  "relatedTopicsTop": [],
  "relatedTopicsRising": [],
  "errors": [],
  "exploreUrl": "https://trends.google.com/trends/explore?q=chatgpt,gemini,claude&date=today%2012-m&geo=US",
  "scrapedAt": "2026-11-08T03:05:04Z"
}

Field	Notes
averageInterest	Google's own average for the period (in comparisons), otherwise the mean of the points that have data
peakInterest, peakDate	Highest point in the period
latestInterest, latestDate	Last complete period. The current, partial week or day is flagged isPartial: true and skipped here
interestByRegion[].value	0-100 within this term: 100 is the region where the term is most popular as a share of all searches there
interestByRegionShare[].share	Comparisons only: this term's percentage of the compared terms in that region (the coloured map on the Trends page)
relatedQueriesRising[].value	Percentage growth. isBreakout: true means growth above 5,000% (Google shows "Breakout")
hasData	false when Google has not enough data for the term. These rows are free
errors	Sections that failed after retries. The rest of the row is still returned

One row per data point (outputFormat: "rows")

Each search term gives one summary row (type: "searchTerm", no nested arrays) plus:

• interestOverTime rows: searchTerm, geo, date, time, value, isPartial
• interestByRegion rows: searchTerm, geoCode, geoName, value, share
• relatedQuery rows: searchTerm, rankType (top/rising), rank, query, value, formattedValue, isBreakout
• relatedTopic rows: searchTerm, rankType, rank, topicId, topicTitle, topicType, value

Trending Now rows

{
  "type": "trendingSearch",
  "geo": "US",
  "rank": 1,
  "query": "texas state vs oregon state",
  "approxTraffic": "2000+",
  "approxTrafficNumber": 2000,
  "startedAt": "2026-11-08T02:40:00Z",
  "pictureUrl": "https://encrypted-tbn2.gstatic.com/images?q=...",
  "pictureSource": "OregonLive.com",
  "news": [
    {"title": "Oregon State Beavers vs. Texas State Bobcats: Live updates today", "snippet": null, "url": "https://www.oregonlive.com/...", "source": "OregonLive.com", "pictureUrl": "https://..."}
  ],
  "exploreUrl": "https://trends.google.com/trends/explore?q=texas%20state%20vs%20oregon%20state&date=now%201-d&geo=US",
  "scrapedAt": "2026-11-08T03:05:01Z"
}

The dataset has three views: Search terms, Trending Now and Data points (long format). The key-value store record RUN_SUMMARY lists each comparison's status, how many terms returned data, any errors, and HTTP statistics.

Use from Claude / ChatGPT (MCP)

Run this actor from an AI assistant through Apify's hosted MCP server. This server URL loads only this actor:

https://mcp.apify.com/?tools=tidewell/google-trends-api

• Claude (claude.ai and Claude Desktop): Settings → Connectors → Add custom connector. Name it Apify, paste the URL, then sign in to Apify.
• Claude Code: claude mcp add apify "https://mcp.apify.com/?tools=tidewell/google-trends-api" -t http
• ChatGPT: add a custom MCP server named Apify with the URL above and OAuth. In the desktop app: Settings → MCP servers → Add server → Streamable HTTP.
• Cursor, VS Code and other MCP clients: {"mcpServers": {"apify": {"url": "https://mcp.apify.com/?tools=tidewell/google-trends-api"}}}. Clients without OAuth can add the header Authorization: Bearer <APIFY_TOKEN>.

Then ask in plain language, for example: "Use tidewell/google-trends-api to compare chatgpt, gemini and claude in the US over the last 12 months and tell me which is rising fastest." The actor appears as the tool tidewell--google-trends-api. Runs are billed to your Apify account at the prices below.

Full setup guide, examples and comparisons: plugin.benjaminbritton.co.uk/google-trends-api/

Pricing

Pay per event. You pay only for what comes back:

Event	Price	When
Search term result	$0.003	Once per search term that returned data. A 3-term comparison costs $0.009, whatever the output format or number of sections
Trending search	$0.001	Once per Trending Now row. A country usually returns 10 rows, so $0.01 per country

There is no start fee. Failed comparisons and terms with no data cost nothing. A run with an empty input (no search terms and no countries) returns Trending Now for the US as a sample, which costs $0.01. Set Max cost per run in the run options to cap spending: the actor stops cleanly before a comparison that would go over the limit.

Examples:

• 100 single keywords for one country: $0.30
• 20 comparisons of 5 brands in 3 countries (300 terms): $0.90
• Trending Now for 10 countries every day for a month: about $3

Tips

• Compare on one line. Values are only comparable within one comparison, so put the terms you want side by side on the same line, up to 5.
• Use a topic ID for ambiguous words. "Python" the language and "python" the snake are different topics. Find the ID in the Trends URL after you choose a topic (/m/05z1_) and use it as the term.
• Interest is relative. 100 is the peak for the chosen terms, place and time, not an absolute search count. For absolute volumes, use a keyword-planner tool. Trending Now's approxTraffic is Google's own rough count.
• Short time ranges need popular terms. For now 1-H or now 4-H, low-volume terms often come back with hasData: false. You are not charged for those.
• Big runs: keep Apify Proxy on, and don't set requestDelaySecs below 1. One comparison takes about 5-20 seconds, depending on the number of terms and sections.

FAQ

Is scraping Google Trends legal? What about Google's terms? Google Trends data is public, aggregated and anonymised, and this actor collects no personal data. Google has no official public Trends API, though, and Google's Terms of Service restrict automated access to its services. Using this actor may therefore conflict with Google's terms, even though it only requests the same public data the Trends website shows, and paces its requests. You are responsible for how you use the data and for checking that your use fits your own legal and compliance requirements. If you need a contractual guarantee, apply for Google's official Trends API (in limited alpha since 2025) or use a licensed data provider. This is not legal advice.

Why are related topics sometimes empty? Google often returns an empty related-topics list to automated requests, even when the Trends page shows topics in a browser. When that happens the actor returns empty relatedTopicsTop / relatedTopicsRising lists rather than guessing. Related queries are not affected. If you don't need topics, turn them off for faster comparisons.

Will the numbers match what I see on trends.google.com? Usually yes, for the same terms, geo, time range, category and search type. Google samples its data, so values can move by a few points between requests and from day to day, and very low-volume terms can show 0 one day and data the next. The exploreUrl in each row opens the same view on Google Trends so you can check.

What happens when Google rate-limits? The actor waits and retries with exponential backoff. With Apify Proxy on (the default), it moves to a fresh IP and session. If a request still fails after all retries, that comparison is reported in RUN_SUMMARY and not charged, and the run continues with the next one. If nothing at all comes back, the run fails with a clear message, so a schedule doesn't silently store empty data.

Can I get daily data for more than 9 months? Google returns daily points for ranges up to about 9 months, weekly points up to 5 years, and monthly points beyond that. For a long daily series, run consecutive custom ranges with an overlap and rescale them using the overlapping days.

Can I schedule it? Yes. Use Apify Schedules (for example, Trending Now every morning, or a weekly brand comparison) and connect the dataset to Google Sheets, Slack, a webhook, Make or Zapier.

Something broke or looks wrong? Open an issue on the actor's Issues tab and include the run ID. Google changes Trends from time to time, and fixes are released quickly.

Keywords

google trends api, google trends scraper, google trends python alternative, pytrends alternative, interest over time, interest by region, related queries, rising queries, breakout keywords, related topics, trending searches, trending now, daily trends, keyword trends, youtube trends, google shopping trends, google news trends, compare keywords, seasonality, seo keyword research