npm Package Scraper · Versions, Downloads, Deps & Maintainers
Pricing
from $1.40 / 1,000 package returneds
npm Package Scraper · Versions, Downloads, Deps & Maintainers
npm Package Scraper (npm Scraper) for extracting package metadata, download stats, dependencies, maintainers, license, and registry scores from npmjs.com. Search by keyword, author, maintainer, or package name. Fast, structured npm dataset extraction with no API key.
Pricing
from $1.40 / 1,000 package returneds
Rating
0.0
(0)
Developer
Tarek Etman
Maintained by CommunityActor stats
0
Bookmarked
1
Total users
0
Monthly active users
4 hours ago
Last modified
Categories
Share

npm Package Scraper · Versions, Downloads, Deps & Maintainers
npm Package Scraper and npm scraper that searches the npm registry on npmjs.com and returns one structured row per package: the latest version, weekly and monthly downloads, how many other packages depend on it, its dependencies, its full publish history, licence, maintainers and the registry's own quality, popularity and maintenance scores.
Maintained by reapX. Every row cites the Apify run that produced it — nothing is inferred, modelled or filled in, and a field absent from the source is absent from the row. The extracted archive for this source is browsable at reapx.dev/data/npm-package-scraper/ and mirrored as an open dataset on Hugging Face and Kaggle. Questions: reapxdev@proton.me
What you get per package
| Field | |
|---|---|
packageName | the permanent registry name, stable across runs |
version · description · keywords · license | the listing |
weeklyDownloads · monthlyDownloads · dependentCount | real usage, not stars |
dependencies · dependencyCount · devDependencyCount | what it pulls in |
versionCount · firstPublishedAt · lastPublishedAt | how long it has been maintained |
maintainers · maintainerCount · publisher | who can publish to it |
isDeprecated · hasTypeScriptTypes · nodeEngine | fitness for use |
unpackedSizeBytes · fileCount | install weight |
qualityScore · popularityScore · maintenanceScore | the registry's own 0-1 scores |
repositoryUrl · homepageUrl · npmUrl · retrievedAt |
How to scrape npm packages and audit dependency trees
Put your package.json dependency names into Specific package names. Every row comes
back with maintainerCount, lastPublishedAt, isDeprecated and license, which is the
whole of a first-pass supply-chain review: single-maintainer packages, packages untouched
for years, packages already deprecated, and licences you cannot ship.
Track an organisation's packages
Set Scope to babel, angular, nestjs or your own org and schedule the run. Each run
captures every package under that scope with its current version and download counts, so a
new release or a sudden drop shows up as a diff between runs on the same packageName.
Find the real leaders in a category
Set Keyword to testing, orm, bundler - anything the ecosystem tags - then set
Minimum dependent packages to something like 100. dependentCount is a far harder
signal than downloads: it counts packages that took a build-time bet on the library.
Input configuration
Every field is optional. The prefilled values below are what runs if you press Start without changing anything.
What to pull
Give queries, package names, or one of the shortcuts below.
| Field | Type | Accepts | What it does |
|---|---|---|---|
searchQueries | list | one value per line | Registry search queries, one per line. Plain text works, and so do npm's own qualifiers: keywords:graphql, author:sindresorhus, maintainer:vercel, scope:babel, not:deprecated, is:unstable. Results from every query are merged and deduplicated. Prefilled with ["keywords:cli"]. |
packageNames | list | one value per line | Exact package names to pull, one per line, including scoped names such as @types/node. |
keyword | string | free text | Shortcut for keywords: |
maintainer | string | free text | Shortcut for maintainer: |
scope | string | free text | Shortcut for scope: |
Filters
All optional.
| Field | Type | Accepts | What it does |
|---|---|---|---|
minWeeklyDownloads | integer | 0 to 1000000000 | Keep only packages downloaded at least this many times in the last week. |
minDependents | integer | 0 to 1000000 | Keep only packages that at least this many other registry packages depend on - the clearest signal that a library is load-bearing. |
excludeDeprecated | true/false | checkbox | Drop packages whose latest version carries a deprecation notice. |
Output
Row depth, volume and politeness.
| Field | Type | Accepts | What it does |
|---|---|---|---|
includeVersionDetail | true/false | checkbox | Read each package's full registry record for version count, first and last publish dates, dependencies, tarball size and deprecation. Turn off for a faster, shallower run. Prefilled with true. |
maxPackages | integer | 1 to 5000 | Hard ceiling on the rows this run produces, and therefore on what it costs. You are charged per package returned. Prefilled with 50. |
requestsPerSecond | integer | 1 to 10 | Politeness against the npm registry. Lower this if you see the registry refusing requests. Prefilled with 5. |
Pricing
Pay per event. The primary event is Package returned (package-returned), charged at $0.002 per event on the free plan, falling to $0.001 on the highest tier.
One npm package delivered as a complete row, including downloads, dependents, dependencies and publish history. Blocked sources and empty result sets are never charged.
Other charged events: apify-actor-start (Actor Start).
Minimum charge cap per run: $0.05.
Tiered discounts apply automatically on every paid Apify plan. Platform usage is absorbed by the Actor, so the per-event price is the whole price.
Usage examples
The prefilled run, which is what Start does with nothing changed
{"searchQueries": ["keywords:cli"],"includeVersionDetail": true,"maxPackages": 50,"requestsPerSecond": 5}
Filter by search queries
{"searchQueries": ["keywords:graphql","author:sindresorhus"],"includeVersionDetail": true,"maxPackages": 50,"requestsPerSecond": 5}
Filter by specific package names
{"searchQueries": ["keywords:cli"],"includeVersionDetail": true,"maxPackages": 50,"requestsPerSecond": 5,"packageNames": ["express","@types/node"]}
Filter by keyword
{"searchQueries": ["keywords:cli"],"includeVersionDetail": true,"maxPackages": 50,"requestsPerSecond": 5,"keyword": "testing"}
Filter by maintainer
{"searchQueries": ["keywords:cli"],"includeVersionDetail": true,"maxPackages": 50,"requestsPerSecond": 5,"maintainer": "sindresorhus"}
Output example
One row, exactly as the actor wrote it to the dataset:
{"packageName": "commander","version": "15.0.0","description": "the complete solution for node.js command-line programs","keywords": ["commander","command","option","parser","cli","argument","args","argv"],"license": "MIT","publisher": "abetomo","maintainers": ["shadowspawn","abetomo"],"maintainerCount": 2,"weeklyDownloads": 474004587,"monthlyDownloads": 1853856927,"dependentCount": 144047,"dependencies": [],"dependencyCount": 0,"devDependencyCount": 9,"versionCount": 124,"firstPublishedAt": "2011-08-14T22:17:51.639Z","lastPublishedAt": "2026-05-29T09:16:23.183Z","isDeprecated": false}
That row carries 29 fields in total; 18 are shown.
Fields on every row
| Field | Type | What it is |
|---|---|---|
packageName | string | The package's permanent name on the npm registry. Unique, stable across runs, and the right key to join runs together over time. |
version | string | The version currently tagged latest. |
description | string | The package's own one-line summary. |
keywords | list | The keywords the package declares, used by registry search. |
license | string | The SPDX licence identifier the package declares. |
publisher | string | The npm user who published the latest version. |
maintainers | list | Every npm user with publish rights on the package. |
maintainerCount | integer | How many maintainers the package has - a one-maintainer dependency is a different risk from a ten-maintainer one. |
weeklyDownloads | integer | Downloads in the last full week, as reported by the registry. |
monthlyDownloads | integer | Downloads in the last full month, as reported by the registry. |
dependentCount | integer | How many other registry packages depend on this one. |
dependencies | list | The runtime dependencies the latest version declares. |
dependencyCount | integer | How many runtime dependencies the latest version pulls in. |
devDependencyCount | integer | How many development dependencies the latest version declares. |
versionCount | integer | How many versions have ever been published. |
firstPublishedAt | string | When the package first appeared on the registry, in UTC. |
lastPublishedAt | string | When the package was last published to, in UTC - the freshest maintenance signal there is. |
isDeprecated | true/false | True when the latest version carries a deprecation notice. |
hasTypeScriptTypes | true/false | True when the latest version declares its own type definitions. |
nodeEngine | string | The Node.js versions the latest release supports. |
unpackedSizeBytes | integer | How large the latest tarball is once unpacked. |
fileCount | integer | How many files the latest tarball contains. |
qualityScore | number | The registry's own quality component, 0 to 1. |
popularityScore | number | The registry's own popularity component, 0 to 1. |
maintenanceScore | number | The registry's own maintenance component, 0 to 1. |
repositoryUrl | string | The package's source repository, normalised to an https URL. |
homepageUrl | string | The homepage the package declares. |
npmUrl | string | The package's page on npmjs.com. |
retrievedAt | string | When this row was read from the registry, in UTC. |
3 named dataset views ship with it: Packages, Supply chain, Health scores. They drive the Output tab in Console and the Output block on the Actor's .md page.
FAQ
Does it need an npm token? No. Everything it reads is the public registry, over plain HTTP, with no credential.
How many packages can one run return? Up to 5,000. Registry search pages at 250 per request and the actor pages through it automatically.
Are scoped packages supported? Yes - @types/node, @babel/core and so on, both in
search results and as explicit names.
What is dependentCount? How many other packages on the registry declare a dependency
on this one. It is the registry's own figure, not an estimate.
Why is version detail optional? It costs one extra registry request per package. Turn it off when you want a fast, shallow sweep of many packages, and on when you want publish history and dependency lists.
What is the stable identifier? packageName. It is the package's permanent registry
address and the right key to join runs together over time.
Unofficial - not affiliated with npm, Inc. or GitHub. Collects public data only. reapx. Contact reapxdev@proton.me.
The full published archive
- npm package scraper archive — every entity this Actor has observed
- All reapx datasets — 20 sources, tens of thousands of pages
- llms.txt · feed.json · feed.xml — the machine-readable index, for agents
Each archive page carries Dataset JSON-LD, a canonical URL and the identifiers used, so an agent can resolve an entity without running anything. Nothing on those pages is estimated or modelled.
🧪 Example input
A real, runnable configuration — this is an actual input this Actor has run with.
{"searchQueries": ["@types"],"includeVersionDetail": true,"requestsPerSecond": 5,"maxPackages": 250}
📄 Sample output
One real row from a real run of this Actor, unedited.
{"packageName": "undici-types","version": "8.9.0","description": "A stand-alone types package for Undici","keywords": [],"license": "MIT","publisher": "GitHub Actions","maintainers": ["matteo.collina","ethan_arrowood"],"maintainerCount": 2,"weeklyDownloads": 298359813,"monthlyDownloads": 1135621830,"dependentCount": 5085,"dependencies": [],"dependencyCount": 0,"devDependencyCount": 0,"versionCount": 129,"firstPublishedAt": "2023-09-19T15:09:35.360Z","lastPublishedAt": "2026-07-24T12:55:33.657Z","isDeprecated": false,"hasTypeScriptTypes": true,"nodeEngine": null,"unpackedSizeBytes": 120779,"fileCount": 48,"qualityScore": 1,"popularityScore": 1,"maintenanceScore": 1,"repositoryUrl": "https://github.com/nodejs/undici","homepageUrl": "https://undici.nodejs.org","npmUrl": "https://www.npmjs.com/package/undici-types","retrievedAt": "2026-08-03T11:10:29Z"}
How it works
- You set the filters below, or none at all - every field is optional.
- The Actor calls the public registry.npmjs.org API directly over HTTP. There is no key to obtain, no login and no headless browser, so a run is fast and cannot break on a UI change.
- It pages through the results until your maximum is reached or the source runs out of matches.
- Every row is pushed to the dataset as it is built - never buffered to the end. A run that hits its time limit still returns everything it collected.
- You are charged
package-returnedat $0.002 per row, and nothing for Apify platform usage. The maximum-results field is therefore your cost cap, and Apify Store discounts apply on top.
💬 Your feedback
Found a bug, or need a field this does not return yet? Open an issue on the Actor's Issues tab, or write to reapxdev@proton.me. Bugs get fixed and reasonable field requests get added.
⚠️ Run outcomes and error handling
This Actor reports what happened in the run's status message, and it always keeps whatever it collected. These are the outcomes you can get and what each one means.
| Outcome | What it means |
|---|---|
| Success | Rows were returned and you were charged package-returned at $0.002 per row. |
| No matches | The source returned nothing for your filters. Nothing is charged. Widen the date window or drop a filter. |
| Partial - source refused | The source rate-limited or refused some requests. The affected items are skipped and named in the log, and everything already collected is still pushed. A block never discards a run's work. |
| Oversized record | A single record exceeded the dataset item limit. It is reported and skipped rather than failing the run. |
What is guaranteed either way
- Every row is pushed as it is built, not buffered to the end of the run. Anything that buffers output loses everything to a timeout, a block or a migration; this does not.
- A field absent from the source is absent from the row. Nothing is inferred, modelled or filled in to make a row look complete.
🧭 Ready-made configurations
These are saved, published configurations of this exact Actor. Open one and run it as it is, or copy its input as a starting point - each link is a live page with the full input and its output schema.
There are 49 of these in total, all listed on the Actor's Examples tab.