Homebrew Scraper: Formulae, Casks & Install Analytics

Homebrew Scraper exports Homebrew package data, version, dependencies, licence, caveats, deprecation status, together with Homebrew's official 30-, 90- and 365-day install counts. Almost no other package registry publishes real adoption numbers; Homebrew does, and this Actor puts them in a table.

Agent skill: SKILL.md

https://api.apify.com/v2/key-value-stores/t7YoTxpZEJOWvw4Ug/records/homebrew-formulae-scraper.md

What you get

Output field	Meaning
name, fullName	The name you type after brew install, and its fully-qualified tap name
resourceType	formula (CLI package) or cask (macOS app)
tap	Which tap it comes from, homebrew/core, homebrew/cask, or a third-party tap
description	One-line description
homepage	Upstream project homepage
version	Current stable version
license	SPDX licence expression. Always null for casks, Homebrew does not record licences for them
dependencies	Runtime dependencies. For casks, a flattened depends_on list such as cask:xquartz or macos:>= :big_sur
buildDependencies	Build-time-only dependencies. Always empty for casks
caveats	The post-install message brew prints, if any
deprecated, deprecationReason	Whether the package is deprecated and why
disabled, disableReason	Whether it has been disabled outright and why
installs30d, installs90d, installs365d	Official install counts over each window
installsOnRequest30d/90d/365d	Installs where the user asked for this package by name, excludes installs pulled in as someone else's dependency. null for casks
sourceUrl	The tarball, git URL or .dmg the package installs from
pageUrl	Its page on formulae.brew.sh
scrapedAt	Run timestamp

A RUN_SUMMARY record in the key-value store holds per-run counts, the filters used, and any package that failed.

Why the analytics matter

installs counts every install, including the ones dragged in as a dependency. installsOnRequest counts only the ones a human typed. The gap between them is the interesting signal:

Package	90-day installs	90-day on request	Read
gh	837,346	836,498	Practically everyone installs it deliberately, a genuine end-user tool
openssl@3	1,440,419	182,675	Mostly a transitive dependency; the real user base is ~13% of the headline number

Ranking developer tools on the headline number alone overstates infrastructure libraries by an order of magnitude.

Common use cases

Measure real developer-tool adoption on macOS. Pull the top of the index by install count and diff it month over month.

{
 "formulae": [],
 "resourceType": "both",
 "maxResults": 250,
 "includeAnalytics": true,
 "includeDependencies": false
}

Track dependency chains for CLI tools. Fetch a named set and read dependencies / buildDependencies to build the graph.

{
 "formulae": ["ffmpeg", "imagemagick", "pandoc", "node", "python@3.14"],
 "resourceType": "formula",
 "includeDependencies": true
}

Spot deprecated packages in a fleet. Feed the output of brew list straight in and filter the dataset on deprecated or disabled.

{
 "formulae": ["wget", "openssl@1.1", "python@3.8", "node", "ffmpeg"],
 "resourceType": "both",
 "includeAnalytics": false
}

Quick start

Three packages, everything about them:

{
 "formulae": ["wget", "ffmpeg", "node"]
}

The 100 most-installed macOS apps:

{
 "formulae": [],
 "resourceType": "cask",
 "maxResults": 100
}

The complete Homebrew index, about 8,500 formulae:

{
 "formulae": [],
 "resourceType": "formula",
 "maxResults": 0
}

Input

Field	Type	Default	Notes
formulae	array	[]	Package names, or full formulae.brew.sh URLs. Empty means index mode, the whole catalogue.
resourceType	string	"formula"	formula, cask or both.
includeAnalytics	boolean	true	Adds the six install-count fields.
includeDependencies	boolean	true	Adds dependencies and buildDependencies.
maxResults	integer	0	Cap per resource type. 0 = no limit.

Combinations that make sense together:

• formulae filled + resourceType: "both": each name is looked up as a formula first, then as a cask. This is the right setting when your list came from brew list and mixes the two.
• formulae: [] + maxResults + includeAnalytics: true: index mode sorts by 90-day installs before capping, so this is how you get "the top N", not "the first N alphabetically".
• formulae: [] + includeDependencies: false: the fastest way to snapshot the whole catalogue for version tracking.
• resourceType: "both" in index mode downloads both dumps (~48 MB combined) and saves up to maxResults of each.

Output example

{
 "name": "wget",
 "fullName": "wget",
 "resourceType": "formula",
 "tap": "homebrew/core",
 "description": "Internet file retriever",
 "homepage": "https://www.gnu.org/software/wget/",
 "version": "1.25.0",
 "license": "GPL-3.0-or-later",
 "caveats": null,
 "deprecated": false,
 "deprecationReason": null,
 "disabled": false,
 "disableReason": null,
 "sourceUrl": "https://ftpmirror.gnu.org/gnu/wget/wget-1.25.0.tar.gz",
 "pageUrl": "https://formulae.brew.sh/formula/wget",
 "dependencies": ["libidn2", "libpsl", "openssl@3", "gettext", "libunistring"],
 "buildDependencies": ["pkgconf"],
 "installs30d": 17456,
 "installs90d": 59779,
 "installs365d": 337567,
 "installsOnRequest30d": 17178,
 "installsOnRequest90d": 58821,
 "installsOnRequest365d": 333559,
 "scrapedAt": "2026-08-06T11:41:20.844Z"
}

RUN_SUMMARY looks like this:

{
 "mode": "named",
 "resourceType": "both",
 "requested": 4,
 "failed": 1,
 "failures": [{ "name": "nope-not-real", "error": "cask \"nope-not-real\" not found (404)" }],
 "packagesSaved": 3,
 "filters": { "resourceType": "both", "maxResults": 0, "includeAnalytics": true, "includeDependencies": true },
 "finishedAt": "2026-08-06T11:41:20.900Z"
}

Limits and behaviour

• Two very different modes. With names, each package is one small request and analytics arrive inline. Without names, the Actor downloads the complete formula.json (~31 MB) or cask.json (~17 MB) dump. That dump is only fetched in index mode, naming packages never pulls it.
• The bulk dumps carry no analytics. Homebrew publishes install counts only on the per-package endpoints and on separate leaderboard files. In index mode the Actor fetches those leaderboards (analytics/install/{30d,90d,365d}.json and analytics/install-on-request/…) and joins them to the dump by name, so installs90d is populated either way.
• Install figures are totalled across variant commands. Homebrew records variant install commands (wget --HEAD) as their own rows on the per-package endpoint, while the leaderboards used by index mode publish one already-summed row per package. The Actor sums the per-package rows to match, so an installs90d from a named run and one from an index run mean the same thing and can be diffed against each other.
• Casks have no installsOnRequest. Homebrew does not publish an on-request leaderboard for casks, so those three fields are null rather than guessed.
• A missing package never aborts the run. An unknown name returns HTTP 404, is recorded in RUN_SUMMARY.failures, and the run continues. The Actor only throws if nothing at all was saved.
• Transient errors are retried. 429 and 5xx get three attempts with linear backoff; a 404 is fatal and not retried.
• Named lookups run 8 at a time, so a list of 200 packages completes in a few seconds.
• Analytics lag by design. Homebrew aggregates daily and the windows are trailing, so a package released last week has small numbers regardless of its popularity.

Finding a package name

It is exactly what you type after brew install. If you are not sure, the name is the last path segment of its page:

Page you see	Name
formulae.brew.sh/formula/wget	wget
formulae.brew.sh/cask/firefox	firefox
formulae.brew.sh/formula/python@3.14	python@3.14

You can paste the whole URL, the Actor extracts the name itself. On your own machine, brew list --formula and brew list --cask print exactly the values this Actor expects.

API example

curl -X POST "https://api.apify.com/v2/acts/arman-bd~homebrew-formulae-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
 -H "Content-Type: application/json" \
 -d '{
 "formulae": ["wget", "ffmpeg", "node"],
 "resourceType": "formula",
 "includeAnalytics": true
 }'

JavaScript example

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_TOKEN' });
const run = await client.actor('arman-bd/homebrew-formulae-scraper').call({
 formulae: [],
 resourceType: 'cask',
 maxResults: 50,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const p of items) console.log(`${p.name} ${p.version}, ${p.installs90d.toLocaleString()} installs/90d`);

Defaults: 1 GB memory, 15 min timeout.

FAQ

Do I need a proxy? No. Proxy configuration is not required to run this Actor.

Do I need a Homebrew account or token? No. You supply no credentials.

What happens if a package is unavailable? A 404 is recorded in RUN_SUMMARY.failures and the run continues with the rest.

Why is license null for my package? It is a cask. Homebrew records licences for formulae only.

Why is installsOnRequest90d null? Same reason, Homebrew publishes on-request analytics for formulae, not casks.

Can I get third-party taps? Yes for named lookups, as long as the tap is indexed on formulae.brew.sh. Index mode covers homebrew/core and homebrew/cask only, because that is what the bulk dumps contain.

Does it tell me if a package is outdated? It gives you the current published version; compare that against what you have installed. The outdated flag in Homebrew's raw JSON refers to the local machine and is meaningless server-side, so it is not exported.

Can I schedule it? Yes, it is designed for scheduled runs. Diff on name plus version for release tracking, or on installs90d for adoption trends.

Can I integrate it with something else? Yes, Apify API, client libraries, webhooks, scheduled runs, dataset exports (JSON/CSV/Excel) or MCP. Output is structured JSON.