SerpApi's Trends product is two engines on the same SERP endpoint. engine=google_trends reprints Google Trends Explore: interest over time, interest by region, related topics, and related queries. engine=google_trends_trending_now reprints Trending Now for a geo and an hours window of 4, 24, 48, or 168. Both stay on Google. Trends API is the exit when the next source is not a Google widget. Request shape is on the API reference.
Two engines, one website
A SerpApi Trends call is a GET. The key sits in api_key. The page is selected with engine. Explore uses q (up to five terms on TIMESERIES), data_type, geo, gprop, and Google's own date tokens such as today 12-m. Trending Now uses geo and hours, not a keyword.
That split matches the website. It does not match a pipeline that already stores source plus keyword plus a date window. The caller still has to learn two engines, then map widget JSON into whatever the warehouse uses.
Related queries are Explore, not a second platform
Related queries and related topics are the reason many teams pick SerpApi over a thin time-series client. Those tables are Google Explore. They are not TikTok sounds, Amazon ASINs, or Steam game names.
Trends API does not ship a related-query table. It ships keyword history (get_time_series), a two-date growth percent (get_growth), and a live ranked list (get_top_trends). For Google history the source string is google search. For the live Google board the type string is Google Trends. Those strings are on the sources reference. Side-by-side query params versus mode sit on Trends API vs SerpApi.
engine names are not source strings
Do not send engine=google_trends as a Trends API source. That string is a SerpApi search parameter. History uses source google search. Trending Now on SerpApi is a second engine. Trending Now on Trends API is get_top_trends with type Google Trends. Mixing those names is a 400 on both sides.
SerpApi's interest_over_time.timeline_data is a widget dump. Trends API rows are date plus value on a 0-100 scale after the second parse of body. Those two JSON trees do not map 1:1.
What a SerpApi Trends pull cannot become
SerpApi can keep covering Google Explore and Trending Now. Check current SerpApi pricing for search credits. The stop is every platform that is not a Google Trends page. A second SerpApi engine for a SERP is still a SERP. It is not app downloads, steam, or wikipedia.
Keep SerpApi when the dashboard must look like the Trends website. Switch when the same cron has to plot Google next to another source on one scale. The broader map of exits is on Google Trends API alternatives.