WordPress Plugins Scraper — installs, downloads, reviews and opportunity scores

Extract the whole WordPress.org plugin directory (about 76,000 plugins) or any tag, search, author or list of plugins, with 70+ fields per plugin: active installs, all-time and 7/30-day downloads with growth, rating and star histogram, support threads and how many were resolved, last release, an abandoned flag, business model, tags, contributors, versions, FAQ and changelog. Add daily download series, reviews with the developer's reply, author portfolios and a category snapshot with a published opportunity score to find plugin ideas worth building.

No login, no API key, no browser. Error rows, duplicates and plugins your filters skip are never charged.

What you can extract

outputType	One row per	Highlights
plugins (default)	plugin	70+ fields; downloads yesterday / 7 d / 30 d / previous 30 d and 30-day growth included at no extra cost
plugin-stats	plugin	daily downloads for up to 729 days, today / last week / all-time totals, share of sites on each version
reviews	review	stars, title, full text, date, public author, replies, developer reply (role, text and delay in hours)
category-snapshot	tag or search term	demand, ratings, staleness, new entrants, commercial share, opportunityScore, abandoned popular plugins, low-rated popular plugins
authors	developer	every plugin they publish, summed installs and downloads, weighted rating, abandoned and commercial counts, member since

Find plugin ideas worth building

• Abandoned plugins that many sites still run: abandonedOnly: true + minActiveInstalls: 100000. Sites need a maintained replacement.
• Popular plugins users dislike: maxRating: 3.5 + minNumRatings: 20 + minActiveInstalls: 10000. Then pull their 1–2 star reviews to read exactly what is missing.
• Category gaps: category-snapshot for tags such as woocommerce, booking, membership, payment and compare their opportunity scores; each row lists the abandoned and low-rated popular plugins of that category.
• Momentum: sort your export by downloadsGrowth30d to see what is gaining installs now.

The opportunity score (published formula)

Every category-snapshot row carries opportunityScore (0–100) and its four components:

opportunityScore = 100 × (0.40 × demand + 0.25 × dissatisfaction + 0.20 × staleness + 0.15 × openness)

• demand = log10(1 + total active installs of the category) ÷ demandLogCap, capped at 1. demandLogCap is 8: 100 million summed active installs saturate demand.
• dissatisfaction = share of rated plugins with a rating below 4.
• staleness = share of plugins without a release in more than 730 days.
• openness = 1 ÷ (1 + populationCount ÷ 50): fewer competitors, more room. populationCount is the size of the whole category as WordPress.org announces it (resultsAnnounced), so a big tag is not made to look open just because fewer plugins were analyzed (formula v1.1; opennessBasis says category, or sample if the category size was unknown).

Which plugins are measured: a snapshot reads the category's listing most installed first (tags) or most relevant first (searches) and aggregates up to maxPluginsPerCategory plugins (default 1,000, up to 10,000). Demand, dissatisfaction and staleness are computed on those pluginsAnalyzed; coverage = pluginsAnalyzed ÷ resultsAnnounced tells you how much of the category that is. For big tags (the woocommerce tag announces 9,761 plugins) raise maxPluginsPerCategory if you want the long tail in the shares; compare categories with similar coverage.

It is deterministic: the same plugins give the same score in every run, so you can compare categories and track one over time. It is arithmetic on public data, not AI. Active installs are the lower bounds of the buckets WordPress.org publishes.

Input examples

Default input {}: the 20 most installed plugins with full records and 30-day downloads.

{ "tags": ["woocommerce"], "abandonedOnly": true, "minActiveInstalls": 10000, "maxItems": 50 }

{ "outputType": "category-snapshot", "tags": ["booking", "membership", "payment", "forms"] }

{ "outputType": "reviews", "pluginUrls": ["https://wordpress.org/plugins/wordpress-seo/"], "reviewsStars": ["1", "2"], "maxReviewsPerPlugin": 100 }

{ "outputType": "plugin-stats", "pluginUrls": ["akismet", "classic-editor"], "statsDays": 365 }

{ "browse": "all", "maxItems": 100000 }

Pasted plugin links are always delivered, whatever the filters. pluginUrls also accepts bare slugs (wordpress-seo).

How each filter works (measured on 17 Oct 2026)

• search, tags, authors and browse are sent to WordPress.org. One listing per tag or term; a search term inside a tag narrows natively. WordPress.org ignores the author when a search term is present, so the actor applies the author itself (author or contributor).
• A search returns the ~1,300 most relevant results (the term "seo" announces 8,644 and serves 1,271). Combine terms with tags to go deeper.
• The popular order repeats some plugins at page boundaries; the actor walks the listing again with different page sizes and recovers them (the woocommerce tag goes from 77% to 99% of the 9,761 announced plugins). browse: "all" walks the whole directory in a stable order.
• Installs, rating, number of ratings, dates, "tested up to", business model and abandoned are checked by the actor on each plugin of the listing, before any extra request. With any of them on, a run checks up to 100 plugins per requested row (at least 2,000); the status message shows how many were checked and skipped. Skipped plugins are never charged.

Output fields (plugin row)

Share of the whole directory with a value (n = 76,015 plugins, 17 Oct 2026):

Field	Filled	Field	Filled
name, slug, url, version, short description	100%	rating (plugins with ratings)	40.9%
activeInstalls + bucket + upper bound	100%	numRatings, ratings1–5	100%
downloaded (all time)	100%	supportThreads, resolved, unresolved	100%
lastUpdated, daysSinceUpdate, abandoned, added	100%	testedUpTo	98.4%
author / authorUsername / contributors	99.6–100%	requiresWp	96.9%
tags	98.2%	requiresPhp	59.0%
descriptionText	99.8%	installationText	90.3%
changelogText	89.2%	faqText	73.4%
versions	81.1%	homepage	74.7%
screenshots	64.1%	bannerUrl	44.2%
upgradeNotice	39.2%	donateLink	37.3%
requiresPlugins (dependencies)	8.9%	blocks	7.4%
businessModel (declared)	1.2%	iconUrl, downloadLink, supportUrl	100%

Download fields (with includeDownloadStats, the default) were filled in 100% of 1,500 real plugin rows; downloadsGrowth30d and downloadsPrev30d in 67% (plugins younger than 60 days have no previous period). Non-empty fields per plugin row: 54.7 of 73 over the whole directory without the download fields (n = 76,028, 17 Oct 2026); with them (the default) add 3 to 5; in 4,189 rows of 50 mixed test runs on Apify (popular plugins weigh more there) the average was 62.9.

Other row types (real rows, 17 Oct 2026)

Row type	n	Average non-empty fields	Fields below 100%
plugin-stats	300	27.9 of 29	rating 90% (plugins without ratings)
review	580	17.0 of 22	developer reply 52% (reviews the developer answered), reviewer username / topic id / last activity 61% (filled when the review page is opened), participants 5% (review-list pages only)
author	300	24.2 of 25	contributor plugin count 27%, member since 99%
category-snapshot	30 tags	37.4 of 38	abandoned popular plugins 53% and low-rated popular plugins 87% (empty when a category has none)

Compared with other WordPress plugin actors on Apify

Read from each actor's published dataset schema and pricing through the Apify API on 17 Oct 2026 (31 actors that scrape the plugin directory; none had more than 2 users in 30 days):

	This actor	Best of the others
Fields per plugin	55–63 non-empty on average (73 keys; see the table above)	32 declared (the richest schema)
7/30-day downloads and growth	yes, by default	no one
Daily download history (up to 729 days) and version share	yes	no one
Category snapshot with a published opportunity score	yes	one actor has a per-plugin score without a published formula
Abandoned / low-rated popular plugins per category	yes	no one
Reviews with the developer's reply and its delay	yes	reply flag only, at $0.008 per topic
Tag coverage when walking a whole tag (woocommerce, 9,761 plugins)	99% (multi-pass)	single pass: ~77%
Price per plugin	$0.001 → $0.0006 by plan, $0.00001 start	$0.00115 → $0.00028 by plan, $0.005 start
Errors, duplicates, filtered rows	never charged	some charge every lookup attempt

Missing values are always null, never omitted. Dates are ISO 8601 (UTC). Contact details (emails, including obfuscated forms such as "name [at] domain.com" and emails inside links, and phone numbers) are removed from names, descriptions, FAQ, changelog, upgrade notices, screenshot captions and reviews, and no avatars are delivered.

Pricing (pay per event)

Event	Price
Plugin record (primary)	$0.001 per plugin ($1.00 per 1,000), lower on higher Apify plans: $0.00078 Silver, $0.0006 Gold and above
Plugin download stats	$0.001 per plugin
Review	$0.0005 per review
Category snapshot	$0.003 per tag or search term
Author	$0.001 per author
Actor start	$0.00001

Never charged: error rows, duplicates within a run, plugins your filters skip, empty reviews, categories or authors without plugins. Rows served from the cache cost the same and carry cachedAt.

Speed and limits

Measured on Apify on 17 Oct 2026: the default input {} finishes in about 4 seconds; 15,000 plugins take about 80 seconds with 4 GB of memory; the whole directory (76,000 plugins) takes about 35 minutes with the default 512 MB and a few minutes with 4 GB — for big exports, raise the memory (the work is CPU-bound). Reviews go at about 1 request per second per connection.

WordPress.org limits requests per connection (about 15 per second on its API and about 1 per second on its review forum, measured). The actor paces itself, spreads the forum requests across connections and, when a connection is slowed down, moves to another one; it never solves a challenge and never logs in.

FAQ

Is the install count exact? No. WordPress.org only publishes buckets ("300,000+"). You get the lower bound, the label and the upper bound of the bucket.

How fresh is the data? By default a plugin record, download series or review that another run read from WordPress.org in the last 6 hours (maxCacheAgeHours, up to 72) can be served from a shared cache; those rows carry cachedAt (when it was read) and cost the same. Listings and the rest are always read live. Set maxCacheAgeHours: 0 to read everything live. Download stats change once a day.

Does it scrape themes? No, plugins only.

What happens when something fails? You get one error row explaining what failed and what to do, the run stays green when the problem is the input, and nothing is charged for it. The ERRORS record of the run lists them all.

This actor is independent and not affiliated with, endorsed or sponsored by WordPress.org, the WordPress Foundation or Automattic. It reads only public pages and public JSON endpoints.