Google Trends Scraper
Pricing
from $2.00 / 1,000 keyword scrapeds
Scrape Google Trends without the failures. Interest over time, interest by region, rising and breakout queries, keyword comparison, and up to 2,600+ trending searches per region with search volume.
Search terms to analyse. Each is scraped separately unless "Compare keywords" is enabled.
[ "protein powder", "creatine"]Paste explore URLs from the Google Trends UI — the keywords, region, time range, category and search surface are read straight out of the link.
[]Scrape all keywords as one comparison on a shared 0-100 scale, like the Google Trends UI. Maximum 5 keywords.
Each selected type is one extra request per keyword.
[ "interestOverTime", "relatedQueries"]Empty for worldwide. Country code (US, GB, IN), sub-region (US-CA), or metro (US-NY-501).
One of now 1-H, now 4-H, now 1-d, now 7-d, today 1-m, today 3-m, today 12-m, today 5-y, all, or a custom range like 2024-01-01 2024-06-30.
Google Trends category id. 0 means all categories (e.g. 71 = Food & Drink, 7 = Finance, 5 = Computers & Electronics).
Which Google property to measure interest on.
Optional. Also return current trending searches for these country codes — with search volume, growth, category and start time. News articles are attached where Google publishes them.
[][ "US", "GB"]Look-back window for trending searches. The Trends UI offers 4, 24, 48 and 168 (7 days); longer windows return far more results — roughly 80 for 4h, 500 for 24h and 2,600 for 168h.
Keep only these Google trending categories. Leave empty for all. Each result carries its own categoryId so you can discover the ids that matter to you.
[]Cap results per trending region.
Fetches Google's trending RSS feed to attach news articles. That feed is a different, much shorter list, so only a few results will match — leave off unless you specifically need article links.
Stop after this many dataset items. 0 means no limit.
Interface language for returned labels, e.g. en-US, de, es.
Parallel keyword jobs. Requests are rate-limited globally, so the delay below is the real throughput control; concurrency mainly helps overlap parsing.
Raise this if you see rate-limit warnings in the log.