trendspy is a Python import, not a hosted API. PyPI describes it as a library for Google Trends. The class is Trends. Methods copy the website: interest_over_time, interest_by_region, related_queries, related_topics, trending_now, trending_now_by_rss. This session's PyPI check showed version 0.1.6, uploaded 25 Dec 2024. Recheck the project page. Trends API is the exit when the job leaves the laptop. The POST is on the API reference.

Methods that mirror the website

from trendspy import Trends
tr = Trends()
df = tr.interest_over_time(['python', 'javascript'])

That call is Google Explore from the machine that ran it. Geo, category, and timeframe tokens are Google's. Multirange compare is a Google feature the package wraps. None of those tokens are Trends API percent_growth values.

trending_now_showcase_timeline independently normalizes up to hundreds of keywords. That is a batch Explore trick. It is not get_top_trends. It is not a 0-100 series that can sit next to source=tiktok.

The wire is still Google's

The readme disclaimer says the library is not affiliated with Google. The same unofficial endpoints that archived pytrends used are the ones a successor has to chase. A maintained wrapper is better than an archived one. It is not a second source.

Trends API source strings include google search, youtube, tiktok, amazon, steam, wikipedia, app downloads, and others. Exact spellings are on the sources reference. Side-by-side of Trends() versus mode sits on Trends API vs TrendSpy.

{
  "mode": "get_time_series",
  "source": "google search",
  "keyword": "python"
}

Parse the body string. No Trends() session. No cookie dance on the cron host.

Last release is a fact to recheck

0.1.6 on 25 Dec 2024 was the latest row on PyPI in this session. A quiet package can still work. A quiet package can also miss the next Google HTML change. Do not treat a last-upload date as a support contract.

When a local class is the wrong runtime

Keep trendspy for exploratory notebooks and RSS-backed trending lists. Leave for anything that must share one key across languages. The other Python successor, with CLI and MCP, is trendspyg. A shopping list of all pytrends exits is the modern-replacement page.