See reports

How to Check if a Python (PyPI) Package Is Safe to Depend On

Run the PyPI Package Intelligence actor on Apify with a list of Python package names. For each one it pulls public data from the PyPI JSON API and pypistats.org - download volume, dependency count, Python versions, license, maintainers and development status - then has OpenAI turn that into a maintenance signal, a popularity tier and concrete risk flags. It answers the Python-specific question plainly: is this package active, stale, or effectively abandoned, before it lands in your requirements.txt or pyproject.toml.

Steps
  1. Open the actor and paste package names. Open PyPI Package Intelligence on Apify and fill the 'packages' field with PyPI package names, one per line (e.g. requests, flask, numpy). Version pins and pypi.org/project/... URLs are tolerated and normalised.
  2. Set the OpenAI key for the risk read. The AI layer needs OPENAI_API_KEY set on the actor; the fact-gathering step (downloads, dependencies, license, maintenance) needs no key. Together they produce the full intelligence record.
  3. Run it and read each record. Click Start. Per package you get latestVersion, weekly and monthly downloads, dependenciesCount, declared pythonVersions, license, maintainersCount, developmentStatus and lastReleaseAgeDays.
  4. Read the AI risk fields. Check maintenanceSignal (active/maintained/stale/abandoned), popularityTier (niche/growing/popular/ubiquitous) and riskFlags like 'unmaintained (last release >18mo)', 'marked Inactive', 'yanked release', 'single maintainer' or 'heavy dependency tree'.
  5. Compare candidates and alternatives. When choosing between libraries for the same job, line them up and let download tier, maintenance signal and the suggestedAlternatives field make the choice obvious.
  6. Export for your supply-chain review. Export to JSON, CSV or Excel, or pull via the API, to keep a standing dependency due-diligence sheet for your Python supply-chain security reviews.
Frequently asked questions

Where do the download numbers come from?

From pypistats.org (the official PyPI download stats), with metadata, versions, classifiers, dependencies and license from the PyPI JSON API. Both are public and key-less. If a stats record is missing or rate-limited, the record still ships, just without download numbers.

How is this different from the npm one?

This is Python-specific: it reads PyPI classifiers like developmentStatus (e.g. '7 - Inactive'), declared Python version support, yanked releases and PyPI maintainer counts - signals that only exist in the Python packaging ecosystem.

What does it cost?

Pay-per-event at $0.10 per package successfully analyzed. A package not found on PyPI comes back with an 'error' field and is not charged.

Can it flag abandoned packages across a whole requirements file?

Yes. Paste your full dependency list and it surfaces unmaintained packages, stale projects, yanked releases and heavy dependency trees across the whole stack in one run.

Does it invent facts or alternatives?

No. The AI summary and risk flags are grounded only in the gathered data, and suggestedAlternatives lists only well-known, widely-trusted substitutes for the same job - otherwise it returns an empty list.

Try it now

Run it on Apify — no install, results in minutes.

Open on Apify