Google Trends API: Interest Over Time, Regions, Related Queries

Get Google Trends data for any keyword as clean JSON, CSV or Excel: interest over time, interest by country, state or city, rising and top related queries, and today's trending searches with news links.

Built to finish reliably. Google Trends rate-limits hard, so this actor paces its requests, backs off and retries automatically, and never charges for a keyword it couldn't fetch. In our tests, 25 keywords plus trending searches for 3 countries finished in about 3 minutes with zero failures.

What you get per keyword

Field	What it is
interestOverTime	Timeline of interest (0-100) for the chosen period
summary	Average interest, peak date, latest value and the change from the first to the last quarter of the period, so you can rank keywords without a spreadsheet
interestByRegion	Interest per country, or per state, city or US metro when you set a location
relatedQueries	Top and rising related searches, with "Breakout" flags
relatedTopics	Top and rising related topics (optional; Google returns these only for some keywords)

Plus, optionally, Trending now: today's top trending searches for any countries you list, with approximate traffic, start time and the news stories behind each one.

Use cases

• SEO and content planning: find rising queries before they peak, and compare topics on one 0-100 scale.
• Product and market research: seasonality, regional demand, and whether a category is growing or shrinking.
• Investing and e-commerce signals: track brand, product or ticker interest daily on a schedule.
• Newsrooms and social media: pull trending searches every hour with the stories behind them.
• AI agents and dashboards: compact, stable JSON with a ready-made summary per keyword.

Input

{
  "keywords": ["air fryer", "pressure cooker", "slow cooker"],
  "compare": true,
  "geo": "US",
  "timeRange": "today 5-y",
  "includeInterestByRegion": true,
  "regionResolution": "REGION",
  "trendingNowGeos": ["US", "GB"]
}

• Keywords: any search terms. Each one gets its own result, or turn on Compare to put up to 5 on the same scale (more are compared in groups of 5).
• Location: a country or region code such as US, GB, DE or US-CA. Empty means worldwide.
• Time range: past hour to 2004-present. A custom range like 2024-01-01 2024-12-31 works through the API.
• Category and search type: narrow to a Google Trends category, or measure Image, News, Shopping or YouTube search instead of web search.

Output example

{
  "type": "keyword",
  "keyword": "bitcoin",
  "geo": "US",
  "timeRange": "today 12-m",
  "summary": { "averageInterest": 34.1, "peakDate": "2026-02-01", "latestValue": 15, "changeFirstToLastQuarterPct": -48 },
  "interestOverTime": [{ "date": "2025-10-12", "time": 1760227200, "value": 41 }],
  "interestByRegion": [{ "geoCode": "US-WY", "geoName": "Wyoming", "value": 100 }],
  "relatedQueries": {
    "top": [{ "query": "bitcoin price", "value": 100, "formattedValue": "100" }],
    "rising": [{ "query": "how to buy bitcoin safely", "value": 10050, "formattedValue": "Breakout" }]
  },
  "scrapedAt": "2026-10-14T13:06:05.921Z"
}

Trending searches come as separate items with type: "trendingNow", rank, title, approxTraffic, startedAt and news.

Pricing

• $4 per 1,000 keywords analyzed (all selected data for that keyword is included).
• $1 per 1,000 trending searches.
• Keywords that fail are free. Set a maximum cost per run and the actor stops cleanly at the limit.

Tips

• Google Trends values are relative (0-100 within your query), not search volumes. Compare keywords in one run to see them on the same scale.
• Short periods (past hour, past day) are noisier; past 12 months or 5 years suits seasonality.
• Schedule a daily run with the same keywords to build your own history.
• A keyword with too little search interest returns an empty timeline; that's Google's data, not an error.

FAQ

Is this legal? It reads the same public, aggregated, anonymous data that anyone sees on trends.google.com. No personal data is collected.

Why do some runs start slower? Google sometimes rate-limits the first request. The actor waits and retries rather than failing.

Missing a feature? Request it in the Issues tab.

About Siftwright

Siftwright is a small data-tools studio. Tools are built and maintained with AI-powered tooling and monitored daily.