Google Trends Scraper — Interest Over Time, Regions & Related Queries

Get Google Trends data as a fast, reliable API: interest over time, interest by region, related queries, and rising topics for any search term — worldwide or for any country, state, or metro area, across Web, YouTube, News, Images, and Shopping search.

No browser. No pytrends. No 429 roulette. This Actor speaks directly to the same JSON endpoints the Trends website uses, with automatic session rotation and backoff when Google rate-limits — so runs are fast (seconds per term, not minutes) and cheap.

Why this exists

• Google Trends has no official public API (the long-promised one is still in limited alpha).
• pytrends, the Python library everyone used for a decade, was abandoned in April 2025 and no longer works reliably.
• Browser-based Trends scrapers are slow, expensive, and flaky — a headless Chrome per keyword to fetch what is ultimately a JSON response.

This Actor is the missing piece: pytrends' functionality, maintained, as a pay-per-result API.

What you get per search term

• 📈 Interest over time — the full 0–100 series at Google's finest granularity for your timeframe (hourly/daily/weekly), with partial-period flags, plus computed average_interest and peak
• 🗺️ Interest by region — countries, states/regions, cities, or US metro areas (DMA)
• 🔍 Related queries — top and rising (including "Breakout" markers — the early-signal goldmine)
• 🏷️ Related topics — top and rising topic entities
• ⚖️ True compare mode — up to 5 terms normalized on one shared 0–100 scale, exactly like the Trends UI comparison (independent scaling when off)
• 🌐 All five search properties — Web, YouTube, News, Images, Shopping

Example output

{
  "term": "chatgpt",
  "geo": "US",
  "timeframe": "today 12-m",
  "property": "web",
  "average_interest": 78.9,
  "peak": { "date": "Oct 19, 2025", "value": 100 },
  "interest_over_time": [
    { "date": "Jul 7, 2025", "timestamp": 1751846400, "value": 71, "is_partial": false }
  ],
  "interest_by_region": [
    { "geo_code": "US-UT", "geo_name": "Utah", "value": 100 }
  ],
  "related_queries": {
    "top": [ { "query": "chatgpt login", "value": 100 } ],
    "rising": [ { "query": "chatgpt agents", "formatted_value": "Breakout" } ]
  }
}

Input

{
  "searchTerms": ["standing desk", "walking pad"],
  "geo": "US",
  "timeframe": "today 12-m",
  "compareAllTerms": true,
  "dataTypes": ["interestOverTime", "relatedQueries"]
}

Timeframes: now 1-H → today 5-y, all, or custom ranges like 2025-01-01 2025-06-30. Geo: country (US), region (US-TX), or empty for worldwide.

Who uses this

• SEO & content teams — seasonality, topic momentum, and rising queries before they're competitive
• E-commerce & product research — validate demand and spot breakout products by region
• Investors & analysts — search interest as an alternative data signal
• Journalists & researchers — quantify attention around events and narratives
• AI agents — "is interest in X growing?" answered with real data via Apify's MCP integration

FAQ

Is there an official Google Trends API? No — Google's announced API remains limited-access alpha. The Trends website is the only public source, and this Actor reads the same JSON its charts read.

Is this a pytrends replacement? Yes, deliberately. pytrends was archived in April 2025. This Actor covers its core surface (interest over time, interest by region, related queries/topics, categories, properties, compare mode) without you maintaining cookies, retries, or proxies.

What do the 0–100 values mean? Google normalizes each request: 100 = peak popularity within your timeframe/geo/term-set. Values are relative, not absolute search volumes. Use compareAllTerms to make multiple terms share one scale.

How does it handle Google's rate limits? Polite pacing, exponential backoff, automatic cookie-session rotation, and per-session proxy identities via Apify Proxy. Failed sub-requests are reported per-term in an errors field rather than failing your whole run.

Can I export CSV/Excel? Yes — every run's dataset downloads as JSON, CSV, XLSX, or XML, and is reachable via API, webhooks, and integrations (Google Sheets, Make, Zapier, LangChain).

Is this legal? It reads the same publicly available, aggregated, anonymized statistics Google publishes to every visitor at trends.google.com.

Roadmap

• Daily trending searches / real-time trends endpoint
• Long-format output option (one row per data point, Sheets-ready)
• Autocomplete/topic-entity resolution (search "mercury" → pick planet vs element)

Found a term or geo that misbehaves? Open an issue — fixes ship fast.