Google Trends data
pytrends is dead. Here's the replacement that works.
If your pytrends scripts started returning 429s and empty responses, it's not you — Google broke the endpoints the library depended on, and the project is unmaintained. This is a drop-in replacement for the data, not the library: the same interest-over-time, related-queries, related-topics, and region breakdowns, served as JSON from a maintained scraper that gets fixed when Google changes things.
Get Google Trends data — $0.004 per resultpytrends vs. this, honestly
pytrends was free; this costs money ($0.004/result — a heavy research session costs cents). What you get for that: it works today, it returns the same surfaces pytrends did, nobody has to babysit proxies or rotate user agents, and breakage is someone else's same-day problem. If you only need one manual lookup, use trends.google.com — it's free. This is for people who need Trends data in a pipeline: SEO tooling, content research, market monitoring, academic work.
From Python, in six lines
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("4l3c/google-trends-scraper").call(
run_input={"searchTerms": ["ai agents"], "timeframe": "today 12-m"})
items = client.dataset(run["defaultDatasetId"]).list_items().itemsFAQ
- Why did pytrends stop working?
- Google changed the private endpoints pytrends relied on, and the library is no longer maintained — requests now return 429s and empty payloads regardless of rate limiting. It isn't your code.
- What does this return that pytrends returned?
- The same four core surfaces: interest over time, interest by region, related queries, and related topics — as clean JSON, with multi-keyword comparison, timeframe and geo parameters.
- How do I call it from Python?
- Via the Apify Python client: pip install apify-client, then call the actor with your keywords and read results from the dataset — about 6 lines of code. Example on the store page.
- What does it cost?
- $0.004 per result row. A typical interest-over-time query for one keyword costs about half a cent. No subscription, no proxy fees.
- Why does this keep working when pytrends broke?
- It's maintained commercially — when Google changes something, fixing it the same day is the product. Runs on Apify's infrastructure with managed proxies, so there are no datacenter-IP blocks to fight.