Similarweb's API v5 is website and app intelligence. A first-request example GETs /v5/website-analysis/websites/traffic-and-engagement with domain (no www.), metrics, granularity, and a date span. The header is api-key. Docs price that endpoint at one credit per metric per result, up to eight if every metric is requested. Rows are visits and engagement, not a Google Trends index. Trends API is the keyword-and-board exit. Sources are on the sources reference.
Domain visits, not search interest
The primary key is a host (nike.com), not a phrase. Metrics include visits, bounce_rate, average_visit_duration, pages_per_visit, unique_visitors. Companion /describe endpoints tell the account which countries and dates it can see. That is a traffic cube.
A Google Trends series keys off a keyword. source google search and keyword=nike is search interest in the word, not sessions on nike.com. Joining the two without a label is a bad chart.
Search Intelligence is still not a 0-100 index
Similarweb also ships keyword endpoints: Keywords Overview (volume, clicks, difficulty, CPC, intent, AI Overview) and SERP Players (which domains collect clicks for a phrase, including clicks over time). Those answers are "who wins the SERP" and "how many clicks." They are not value 0-100 on google search.
REST is for a few domains now. Batch is for large lists as CSV. Trends API has no Batch CSV job. It has three modes. Parse the body string.
Top Websites is the closer feed
Trends API live type Top Websites needs a category. Official category names sit on category feeds. The result is a ranked site list, not bounce rate.
{
"mode": "get_top_trends",
"type": "Top Websites",
"category": "eCommerce",
"limit": 10
}
That is still not Similarweb's traffic-and-engagement table. Keep Similarweb when the KPI is visits or SERP click share. Leave when the KPI is a keyword's 0-100 path on google search or youtube. Planner volume instead of visits is Keywords Everywhere.