Google Trends Scraper: Unlimited Keywords + Daily Data

Google Trends Scraper

Google Trends with the two limits removed:

• No 5-keyword limit. Google compares at most 5 terms, and every comparison has its own 0–100 scale, so numbers from two comparisons can't be put side by side. Put up to 100 keywords on one line and get them all back on one 0–100 scale, ranked.
• Daily data for any time range. Google switches to weekly data after ~9 months and monthly after 5 years. Turn on Daily data and get one point per day for 12 months, 5 years, or 2004 to now.

Plus everything else Google Trends has:

• Explore: interest over time, interest by region and related queries (top + rising) for any keyword, country and time range, or paste any Google Trends URL.
• Trending Now: every search trending in any country right now: search volume, % increase, when it started, whether it's still active, related searches and category. Covers the past 4 hours to 7 days, with hundreds of trends per country.

Built to keep working when Google rate-limits: every exit IP is paced, retired before Google blocks it, and swapped out immediately on a 429, so runs finish instead of hanging.

Compare any number of keywords

Put them on one line, separated by commas:

coffee, tea, matcha, espresso, latte, cold brew, chai, kombucha, hot chocolate, yerba mate, decaf, oat milk

You get one report with every term on the same scale, plus a ranking:

"ranking": [
  { "rank": 1,  "term": "coffee",     "average": 58.935, "vsTop": 1.0,    "lowPrecision": false, "tooSmallToMeasure": false },
  { "rank": 2,  "term": "tea",        "average": 32.324, "vsTop": 0.5485, "lowPrecision": false, "tooSmallToMeasure": false },
  { "rank": 3,  "term": "espresso",   "average": 6.298,  "vsTop": 0.1069, "lowPrecision": false, "tooSmallToMeasure": false },
  { "rank": 12, "term": "yerba mate", "average": 0.398,  "vsTop": 0.0067, "lowPrecision": false, "tooSmallToMeasure": false }
]

(US, past 5 years, real run.) Note yerba mate: in Google's own 5-term comparison with coffee it shows 0 all the way, because Google rounds to whole numbers. Here it has a real value.

How it works. First every term is sized against one shared term. Then terms of similar size are compared in overlapping groups of 5, each group sharing one term with the next, and the groups are chained onto one scale (the anchor method from the G-TAB research paper). Comparing similar-sized terms is what keeps small terms precise.

How accurate. In repeat runs our ratios matched to within 1% (matcha 0.0615 vs 0.0616 of coffee), more consistent than Google's own direct comparison, which moved 7% between identical requests. A term over ~200× smaller than every other term in your list can't be measured next to them by any method: it's marked tooSmallToMeasure (put it in a list with similar-sized terms). Terms whose link to the rest was weak are marked lowPrecision.

Daily data for any time range

Turn on Daily data for any time range and pick 12 months, 5 years, 2004–present or a custom range. You get one point per day ("resolution": "day"), plus Google's own weekly or monthly series in weeklyOrMonthlyInterest, interest by region and related queries.

How it works. The range is fetched in ~8-month windows (Google gives daily data for those), and each window is scaled so its weekly totals match Google's full-range series. Every window is tied to the same full-range reference, so errors don't build up from window to window.

How accurate. Coffee vs tea, US, 5 years (real run): our daily data, added up into weeks, matched Google's own weekly series with 0.99 correlation (tea 0.96). Over the last 90 days it matched Google's own daily numbers with 0.98 correlation, and the tea-to-coffee ratio came out 0.439 vs Google's 0.440.

What people use it for

• SEO and content planning: find rising queries ("Breakout", +1,900%) before they're competitive
• Market research: compare brands, products or categories across countries and states
• Investing and trading signals: track search interest in tickers, products and themes over time
• Keyword research at scale: rank a whole keyword list on one scale instead of in batches of 5
• Forecasting and seasonality: years of daily data for models, to match against daily sales, or to time campaigns and stock
• Local demand: see which regions or cities search for a product most

What you get

Explore: one dataset item per report

{
  "searchTerms": ["coffee", "tea"],
  "geo": "AU",
  "timeRange": "today 12-m",
  "averages": { "coffee": 78, "tea": 41 },
  "interestOverTime": [
    { "date": "2026-09-20T00:00:00+00:00", "label": "Sep 20 – 26, 2026", "values": { "coffee": 81, "tea": 40 }, "isPartial": true }
  ],
  "interestByRegion": [
    { "geoCode": "AU-QLD", "geoName": "Queensland", "values": { "coffee": 70, "tea": 30 }, "hasData": true }
  ],
  "relatedQueries": {
    "coffee": {
      "top": [{ "query": "coffee machine", "value": 100, "formattedValue": "100", "link": "https://trends.google.com/trends/explore?q=coffee+machine&date=today+12-m&geo=AU" }],
      "rising": [{ "query": "how to descale coffee maker", "value": 1900, "formattedValue": "+1,900%", "isBreakout": false }]
    }
  },
  "trendsUrl": "https://trends.google.com/trends/explore?q=coffee%2Ctea&date=today+12-m&hl=en-US&geo=AU",
  "scrapedAt": "2026-09-23T09:12:44+00:00"
}

Values are Google's relative interest, from 0 to 100, where 100 is peak popularity for the terms, place and time you chose. Resolution follows Google: minute-level for the past hour or 4 hours, hourly for 7 days, daily up to 90 days, weekly up to 5 years, and monthly since 2004.

Spreadsheet-friendly CSVs are also saved to the run's key-value store, free: interest_over_time.csv, interest_by_region.csv and related_queries.csv, one row per value, ready for Excel or Google Sheets.

Trending Now: one dataset item per trend

{
  "type": "trendingNow",
  "rank": 1,
  "title": "jaxson dart",
  "geo": "US",
  "windowHours": 168,
  "searchVolume": 500000,
  "increasePercent": 1000,
  "startedAt": "2026-09-22T20:00:00+00:00",
  "endedAt": null,
  "isActive": true,
  "relatedSearches": ["jaxson dart", "jaxson dart injury", "jaxon dart"],
  "categories": ["Sports"],
  "newsArticleCount": 4,
  "trendsUrl": "https://trends.google.com/trends/explore?q=jaxson+dart&geo=US&date=now+7-d"
}

searchVolume is Google's lower bound, e.g. 500000 means "500K+ searches".

How to use it

Explore

1. Add your search terms, one report per line. Put coffee, tea on one line to compare them (any number of terms, up to 100), or paste a Google Trends explore URL.
2. Pick a country (US, GB, AU, US-CA…) or leave it empty for worldwide.
3. Pick a time range, or type a custom one like 2024-01-01 2024-06-30.
4. Run it. Every report lands in the dataset; export it as JSON, CSV or Excel, or call the API.

Optional: daily data for any time range, category, search type (web, images, news, Shopping, YouTube), region resolution (countries, states, cities, US metro areas) and which sections to include.

Trending Now

1. Set What to scrape to Trending Now.
2. Add one or more countries or regions (US, GB, AU, US-CA…).
3. Pick a window (4 hours to 7 days). Optionally filter by category (Sports, Business and Finance, Technology… 19 in all), keep only active trends, or cap how many per country.

Pricing

• Explore: per report, not per row. One report covers everything above for up to 5 compared terms.
• More than 5 keywords or daily data: charged per Google Trends query it takes, at the report price. Typical: 12 keywords ≈ 6 reports, 40 keywords ≈ 20, daily data for 12 months = 3, 5 years = 10, 2004–present ≈ 41. If your maximum cost for the run can't cover a report, it's skipped, not half-done.
• Trending Now: per trend returned.

Failed reports, and terms Google has no data for, are never charged.

Past trending searches

Google only keeps Trending Now for 7 days. For older days, PastCharts has free daily history by country back to November 2024. Use this Actor for today's data.

Tips

• Use residential proxies (the default). Google rate-limits datacenter IPs heavily. Trending Now is light, at one request per country.
• Up to 10 reports run in parallel, each on its own IP.
• Set a maximum cost per run in the run options, and the scraper stops cleanly when it reaches it.

Limits

• Google Trends values are relative (0–100) and sampled, so repeated runs can differ slightly. This is how Google Trends itself behaves.
• Combined comparisons and daily data are estimates built from several Google Trends queries, accurate to the degree shown above, not Google's own published numbers. Google Trends shows relative interest, not search counts. Check anything important before you rely on it (for example, before investment or business decisions).
• Very rare terms may return no data. That's Google's "not enough search volume".
• Comparisons of more than 5 keywords return interest over time and the ranking; for regions and related queries, run those terms in lines of up to 5.
• Daily data works for up to 5 terms per line.
• Related topics are not included. Google now withholds them from every anonymous (logged-out) session and returns an empty list, so any scraper that doesn't log in to a Google account gets none. Related queries are unaffected.