pytrends-reliable — maintained pytrends alternative (Google Trends API for Python)

pytrends is archived (last release April 2023) and many users now hit TooManyRequestsError: The request failed: Google returned a response with code 429, or find that trending_searches() no longer works.

pytrends-reliable keeps the exact pytrends interface — change one import and your code runs again:

# before
from pytrends.request import TrendReq
# after
from pytrends_reliable.request import TrendReq

pytrends = TrendReq(hl='en-US', tz=360)
pytrends.build_payload(['coffee', 'tea'], timeframe='today 3-m', geo='US')

pytrends.interest_over_time()               # DataFrame: date index, one column per keyword, isPartial
pytrends.interest_by_region(resolution='REGION', inc_geo_code=True)
pytrends.related_queries()                  # {'coffee': {'top': df, 'rising': df}, 'tea': {...}}
pytrends.related_topics()
pytrends.trending_searches(pn='united_kingdom')   # works again
pytrends.suggestions('tesla')

pip install pytrends-reliable

What's different from pytrends

Reliable backend

Free direct requests are fine for occasional use, but Google rate-limits IP addresses quickly — especially servers, notebooks in the cloud and scheduled jobs. When that happens you can route the same calls through a managed Google Trends scraper that rotates IP addresses:

pytrends = TrendReq(backend='apify', apify_token='YOUR_APIFY_TOKEN')
# or: export APIFY_TOKEN=... and keep TrendReq() — it switches automatically after repeated 429s

• Same methods, same DataFrames.
• Uses the Google Trends Scraper on Apify: about $1.50 per 1,000 searches; Apify's free plan includes monthly credits.
• Each build_payload + data calls = one search (the data is fetched once and cached).
• Slower per call (a cloud run takes ~20–40 s) but it does not get blocked.

Disclosure: I built both this library and the Apify scraper. The direct backend is free and needs no account; the reliable backend is optional.

Supported methods

Timeframes are Google's: 'today 5-y', 'today 12-m', 'today 3-m', 'today 1-m', 'now 7-d', 'now 1-d', 'now 4-H', 'now 1-H', 'all', '2024-01-01 2024-06-30', '2024-01-01T10 2024-01-03T10'.

Tips against 429 with the direct backend

• Reuse one TrendReq object; avoid many parallel scripts on the same IP.
• Wait a few seconds between build_payload calls.
• TrendReq(max_429_retries=6) retries longer; proxies=['https://...'] rotates your own proxies (same as pytrends).

License

Apache 2.0. Keeps the public interface of pytrends (Apache 2.0, © General Mills). Independent project, not affiliated with Google or the pytrends maintainers.