npm & PyPI Package Metadata API: Releases, Licenses, Deps avatar

npm & PyPI Package Metadata API: Releases, Licenses, Deps

Pricing

from $2.00 / 1,000 package report returneds

Go to Apify Store
npm & PyPI Package Metadata API: Releases, Licenses, Deps

npm & PyPI Package Metadata API: Releases, Licenses, Deps

Package metadata and optional dated release history from official npm and PyPI APIs. Get versions, licenses, dependencies, repositories and deprecation status.

Pricing

from $2.00 / 1,000 package report returneds

Rating

0.0

(0)

Developer

asaf meizner

asaf meizner

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

13 days ago

Last modified

Categories

Security

Limited permissions

Share

npm & PyPI Package Metadata API

Get clean, structured metadata for npm and PyPI packages in one run: latest version, release date, days since the last release, first release, release count, license, summary, repository, runtime requirement, dependency list and deprecation or yank status. Data comes straight from the official public registries (registry.npmjs.org and pypi.org JSON APIs). No login, no scraping of web pages.

Use cases

  • Dependency health checks: flag packages not released for years or marked deprecated.
  • License review: list licenses across a whole dependency list.
  • Release watching: schedule a daily run and compare latestVersion with your last run.
  • Open-source research: compare maintenance activity across ecosystems.

Input

{ "packages": ["npm:express", "npm:@types/node", "pypi:requests", "pypi:numpy"] }

Provide 1–500 packages per run, written as npm:<name> or pypi:<name>. Duplicates are removed. The default input is ["npm:express", "pypi:requests"]; releaseLimit defaults to 0 (no history), and releasedSince is optional.

For a small first run, use the two default packages. If both return reports, the package-report events total $0.004. The four-package example above totals $0.008 if all four return reports. Apify platform usage may also apply. Repeated runs charge again for each report returned.

Output (one row per package)

{
"ecosystem": "npm",
"name": "express",
"status": "ok",
"latestVersion": "5.3.0",
"latestReleaseDate": "2026-10-09T12:56:54.505Z",
"daysSinceLatestRelease": 9,
"firstReleaseDate": "2010-12-29T19:38:25.450Z",
"releaseCount": 290,
"license": "MIT",
"summary": "Fast, unopinionated, minimalist web framework",
"repository": "https://github.com/expressjs/express",
"dependencyCount": 28,
"runtimeRequirement": ">= 18",
"deprecated": null
}

Packages that do not exist return {"status": "not_found"} and are not charged.

Pricing

Pay per event: $0.002 per package report returned (about $2 per 1,000 packages). Not-found rows are free.

Notes

  • Sequential, polite requests with an identifying User-Agent; the run stops on repeated rate-limit or server errors.
  • Only package metadata published by the registries. Maintainer names and e-mail addresses are never returned.
  • For PyPI, optional "extra" dependencies are excluded from the dependency list; the license is blank when the package pastes full license text into the field.

Release history and monitoring

Set releaseLimit to 1–100 to include a releases array in each package report, sorted by publication timestamp (not semantic version). Default 0 preserves the original output. Optionally set releasedSince to an ISO timestamp with timezone; only releases strictly after that time appear. This filters history only: package reports with no matching releases are still returned and charged at the existing $0.002 per package, with no extra release event.

{"packages":["npm:express","pypi:requests"],"releaseLimit":10,"releasedSince":"2026-01-01T00:00:00Z"}

npm records include version, releasedAt and deprecation message. PyPI records use the earliest file upload for each version and indicate whether all files are yanked. Undated and empty releases are omitted. Prereleases are included. Registry publication timestamps are not changelog text; removed releases cannot be recovered, and this actor does not save monitoring state. Schedule runs and retain your own comparison timestamp.