Google Trends Scraper — reliable, no browser

Get Google Trends data through an API: interest over time, interest by region, related queries and related topics, for one keyword or for up to 5 keywords compared together.

A pytrends alternative that works in 2026. pytrends was archived in April 2025 and the official Google Trends API is still limited-access alpha. This Actor gives you the same data as the Google Trends website, as JSON, CSV or Excel, from the Apify API, Python, JavaScript, Make, n8n, Zapier or an AI agent (MCP).

Why this Actor

Runs either return data or fail with a clear reason. They don't spin for minutes and return nothing.

• No browser. Each keyword takes about five lightweight HTTP requests, spaced at least two seconds apart. In a test on October 21, 2026, three keywords with all four sections returned 462 rows in 33 seconds.
• Plays by Google's rules. It never loads pages that Google's robots.txt disallows, and it never rotates IPs or sessions to get around a block. If Google answers 429 (too many requests), it slows down (honoring Retry-After, otherwise pausing about 30 seconds) and retries at most twice. If Google answers 403 or keeps answering 429, the run stops and tells you to try again later.
• Bounded time. Server errors are retried with exponential backoff. Each keyword group has a 150-second budget covering all its requests and retries. When the budget runs out, the affected sections are reported as errors.
• A status for every section. Each keyword group and section gets a status (ok, empty or error, with the reason) in the run's key-value store under the key STATUS, so an empty result is never ambiguous. A keyword with no Google Trends data is reported as empty, not as a failed run.
• Flat output. The dataset holds only data rows, one per data point, ready for CSV, Google Sheets or a database.
• You pay only for rows you get. $1.00 per 1,000 data rows. Status records and failed sections are never charged.

What you can use it for

• SEO and content planning: find rising queries ("Breakout") around a topic before competitors write about it.
• Product and market research: compare brands or products over 5 years, by country or region.
• Seasonality: pull weekly interest for a keyword list and feed it to a forecasting model.
• Dashboards and alerts: schedule the Actor daily and push the dataset to Google Sheets, Slack or a database.

Input

Field	Description	Default
searchTerms	Keywords to analyse	—
geo	Country or region code (US, FR, FR-J); empty = worldwide	empty
timeRange	now 7-d, today 1-m, today 3-m, today 12-m, today 5-y, all…	today 12-m
sections	Any of interest_over_time, interest_by_region, related_queries, related_topics	all
compareKeywords	Compare keywords together (groups of 5, values on one shared scale)	false
category, property	Google Trends category ID; web, images, news, YouTube or Shopping search	0, web
hl	Interface language of the labels (dates, topic names)	en-US
maxAttempts	Attempts per request before giving up	5
proxyConfiguration	Apify Proxy settings; one session is used for the whole run	Apify Proxy

Output examples

{"type": "interest_over_time", "keyword": "coffee", "geo": "FR", "timestamp": 1760832000, "date": "Oct 19 – 25, 2025", "value": 26, "hasData": true, "isPartial": false}
{"type": "related_queries", "keyword": "coffee", "metric": "rising", "rank": 1, "query": "coffee crush krups", "value": 79600, "formattedValue": "Breakout", "isBreakout": true}

Status (key-value store, key STATUS):

[{"keywords": ["coffee"], "section": "related_topics", "status": "empty", "rows": 0, "error": null}]

Google Trends values are relative (0–100 within the request), not search volumes.

Call it from Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("gzeff/google-trends-reliable").call(run_input={
    "searchTerms": ["coffee", "tea"],
    "geo": "US",
    "timeRange": "today 12-m",
    "sections": ["interest_over_time", "related_queries"],
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row["type"], row["keyword"], row.get("date") or row.get("query"), row["value"])

FAQ

How much does a typical run cost? One keyword with all four sections returns roughly 100 to 400 rows, so about $0.10 to $0.40. Use sections to request only what you need.

Why did a section come back empty? Google Trends has no data for that keyword, place or period (common for rare keywords or small regions). The STATUS record says empty, and you are not charged.

Why is related_topics almost always empty? Since at least October 2026, Google's own endpoint returns an empty list for related topics, even in a regular browser. The section is kept in case Google restores it. It costs nothing when empty, and you can leave it out of sections.

What happens if Google rate-limits the run? The run stops instead of trying to get around the limit. Keyword groups that were not processed are listed as skipped in STATUS. Rows already collected are kept, and only those are charged.

Can I compare more than 5 keywords? This Actor compares at most 5 keywords per request, like the classic Google Trends Explore view. With compareKeywords on, longer lists are split into groups of 5. Values are comparable within a group, not across groups.

Limits

• Google may rate-limit any automated client. This Actor respects those limits: it slows down, then stops and reports it. It cannot guarantee results for every request.
• For large-scale production use, consider applying to Google's official Trends API (alpha).
• This is an unofficial tool. It is not affiliated with Google.
• Questions and feature requests: open an issue on this Actor.