App Store, Google Play & Steam Reviews Scraper - thousands of reviews, no login

Collect user reviews and rating breakdowns for Apple App Store, Google Play and Steam apps in one clean, unified dataset. Go to tens of thousands of reviews on Google Play and Steam, cover many countries and languages, filter to 1-2 star complaints or keywords like "crash" and "refund", and run it on a schedule to get only the new reviews. No account, no API key.

What you get

Review rows (type: "review"), same fields for every store:

Group	Fields
Identity	store - appId - appName - reviewId - country - language - url
Rating	rating (1-5) - sentiment (positive / neutral / negative) - recommended (Steam)
Content	title (Apple) - text - textLength - wordCount
Who and when	author - date (ISO, UTC) - version (app version reviewed)
Votes	helpfulVotes - funnyVotes, commentCount (Steam)
Developer reply	hasDeveloperReply - developerReply (author, text, date) - Google Play and Steam
Steam extras	playtimeAtReviewHours - playtimeHours - purchasedOnSteam - receivedForFree - earlyAccess - playedOnSteamDeck - authorGamesOwned - authorReviewCount

App rows (type: "app", one per app and country): name - developer - rating - ratingCount - ratingHistogram (1 to 5 stars counts) - version - category - contentRating - price / free - released / updated - icon - description, plus Google Play installs, offersIAP, containsAds, Apple languages, releaseNotes, sizeBytes and Steam reviewScoreDesc, positivePercent, totalPositive, totalNegative, metacriticScore.

Error rows (type: "error"): an app that could not be found or read, with the reason.

Example review row (Google Play)

{
  "type": "review",
  "store": "google-play",
  "appName": "Spotify: Music and Podcasts",
  "reviewId": "3b7d8e89-30c5-4e13-aabd-7c638e4f70bd",
  "country": "de",
  "rating": 1,
  "sentiment": "negative",
  "text": "Seit dem letzten Update stuerzt die App staendig ab ...",
  "author": "Ein Google-Nutzer",
  "date": "2026-10-02T11:08:41.000Z",
  "version": "9.0.12.745",
  "helpfulVotes": 14,
  "hasDeveloperReply": true,
  "developerReply": { "author": "Spotify AB", "text": "Hi! Things should be running smoothly now ...", "date": "2026-10-02T14:22:49.000Z" }
}

How deep can it go?

Store	Depth	Notes
Google Play	tens of thousands (about 200 reviews per request)	newest or most helpful, one star level can be filtered server-side, developer replies included
Steam	unlimited	newest or most helpful, per language, positive / negative filtered server-side, playtime and developer response included
Apple App Store	about 500 per country and ordering, about 1,000 per country with both orderings merged	Apple's public reviews feed is capped. Add countries with extraCountries to go wider. Developer replies are not in this feed

Input examples

Complaints from the last 90 days, three countries, two stores:

{ "apps": ["com.spotify.music", "https://apps.apple.com/us/app/spotify-music-and-podcasts/id324684580"],
  "country": "us", "extraCountries": ["gb", "de"], "maxRating": 2, "daysBack": 90, "maxReviewsPerApp": 3000 }

Competitor benchmark with keywords:

{ "apps": ["com.duolingo", "com.babbel.mobile.android.en", "1245620"], "keywords": ["subscription", "refund", "crash"],
  "maxReviewsPerApp": 1000 }

Daily alert feed, only reviews newer than the previous run:

{ "apps": ["com.whatsapp"], "onlyNewReviews": true, "maxRating": 2, "stateStoreName": "whatsapp-complaints" }

An app is detected as Apple (id310633997 or an apps.apple.com link), Google Play (com.whatsapp or a play.google.com link) or Steam (a store link or a plain appid such as 440). To force a store use {"app": "...", "store": "apple"}.

Use cases

• Product and QA teams - what broke in the last release? Filter by version and 1-2 star ratings.
• Competitor research - compare complaints and praise across rival apps and countries.
• App store optimisation and support - find unanswered reviews (hasDeveloperReply: false).
• Sentiment analysis and LLM pipelines - rows carry sentiment, wordCount and clean text.
• Game studios and publishers - Steam playtime-at-review separates veterans from refunders.
• Investors and analysts - rating histograms and review velocity as traction signals.

FAQ

Is a login or API key needed? No. All three stores are read through their public endpoints.

Why do I only get about 500 Apple reviews per country? Apple limits its public feed to 10 pages. The Actor merges the newest and most-helpful orderings (about 1,000) and you can add more countries for more coverage.

Does Steam rating mean stars? Steam only has recommend / not recommend, so rating is 5 or 1 and recommended shows the real vote. The app row has the overall score (reviewScoreDesc, positivePercent).

Which countries and languages work? Any two-letter country the store supports. Steam languages: english, french, german, spanish, japanese and the rest of Steam's list, or all; ISO codes like fr are accepted too.

How do I avoid duplicates across countries? Reviews are de-duplicated by reviewId within each app.

What if one app fails? It becomes an error row; the others are returned and the failed app is not billed.

What does it cost? Per review delivered (review-scraped) plus per app read successfully (app-scraped).

Output

Dataset rows (review / app / error) and a SUMMARY record in the key-value store with per-app counts, rating and any notes (for example when Apple's feed limit was reached).