Scoop Windows Package Manager Scraper

Scrape app metadata from Scoop Windows package manager buckets. Search for apps by name or fetch specific app manifests across the Main, Extras, Versions, and Java buckets.

Features

• Search Apps — find apps in any Scoop bucket by name substring (e.g., search "git" finds git, git-lfs, git-credential-manager, etc.)
• Get By Name — fetch manifest data for specific apps by exact name
• Covers all major buckets: Main, Extras, Versions, Java
• Extracts version, description, homepage, license, installer URL, binaries, dependencies, and shortcuts
• No authentication required; reads from GitHub raw content

Input

Field	Type	Description	Default
mode	select	searchApps or getByName	searchApps
searchQuery	string	App name filter (case-insensitive substring)	—
bucket	select	Which bucket to search: main, extras, versions, java	main
appNames	array	Exact app names for getByName (e.g. ["git", "nodejs"])	[]
maxItems	integer	Max records to return (1–1000)	50

Example: Search Apps

{
  "mode": "searchApps",
  "searchQuery": "git",
  "bucket": "main",
  "maxItems": 20
}

Example: Get By Name

{
  "mode": "getByName",
  "appNames": ["git", "nodejs", "python", "7zip"],
  "bucket": "main",
  "maxItems": 10
}

Example: Search Extras Bucket

{
  "mode": "searchApps",
  "searchQuery": "vscode",
  "bucket": "extras",
  "maxItems": 5
}

Output

Each record contains:

Field	Description
appName	App identifier (filename without .json, e.g. git)
version	Current version in the bucket
description	Short description of the app
homepage	Official homepage URL
license	License identifier (e.g. MIT, GPL-2.0-only, Freeware)
installerUrl	Primary download URL for the app installer
hash	SHA256 hash of the installer for integrity verification
depends	List of app dependencies
suggest	Suggested companion apps (from manifest)
bin	List of executables/binaries added to PATH
shortcuts	Desktop/Start menu shortcuts created
bucket	Which bucket this app came from
sourceUrl	Link to the app on scoop.sh
recordType	Always "app"
scrapedAt	ISO timestamp when record was scraped

Sample Output

{
  "appName": "git",
  "version": "2.54.0",
  "description": "A free and open source distributed version control system.",
  "homepage": "https://gitforwindows.org",
  "license": "GPL-2.0-only",
  "installerUrl": "https://github.com/git-for-windows/git/releases/download/v2.54.0.windows.1/PortableGit-2.54.0-64-bit.7z.exe#/dl.7z",
  "hash": "bea006a6cc69673f27b1647e84ab3a68e912fbc175ab6320c5987e012897f311",
  "bin": ["bin\\sh.exe", "bin\\git.exe", "git-bash.exe"],
  "shortcuts": [["git-bash.exe", "Git\\Git Bash", "--cd-to-home"]],
  "bucket": "main",
  "sourceUrl": "https://scoop.sh/#/apps?q=git&s=0&d=1&o=true",
  "recordType": "app",
  "scrapedAt": "2026-01-01T12:00:00+00:00"
}

Buckets

Bucket	Contents
main	Stable, widely-used developer tools (git, nodejs, python, 7zip, cmake, etc.)
extras	Additional apps not meeting main criteria (vscode, firefox, chrome, etc.)
versions	Older/specific versions of apps in main
java	Java Development Kits and runtimes (JDK 8, 11, 17, 21, etc.)

Use Cases

• Software inventory — track what apps are available in Scoop and their versions
• Update monitoring — compare installed versions against bucket manifests
• License compliance — audit licenses across your dev toolchain
• CI/CD tooling — build automated scripts that reference latest package versions
• Package discovery — explore what's available before installing

FAQ

Does this require a GitHub token? No. The actor reads from public GitHub repositories without authentication. For high-volume scraping, the GitHub API rate limit is 60 requests/hour unauthenticated.

Why are some fields missing? Fields like depends, bin, shortcuts, and suggest are optional in Scoop manifests and only appear when defined by the package maintainer.

Can I search multiple buckets at once? Currently, one bucket per run. Run the actor multiple times with different bucket values to cover all buckets.

What is the installerUrl? The primary download URL for the portable app archive. For apps with architecture-specific URLs, the 64-bit URL is preferred.

How current is the data? The actor fetches live data from GitHub at scrape time, so results reflect the current state of the bucket.