App Store & Google Play Reviews Scraper — iOS + Android in One Run

Scrape reviews from both app stores in a single run, across as many countries as you like, and get one normalized schema back. Add an app twice — once per store — and you can compare iOS against Android directly.

Most review scrapers do one store, one country. This one does both stores, 150+ countries, only-what-changed runs, and a rules-based insights report with no LLM and no API key.

What you get

Per review

Field	iOS	Android
Rating, review text, author, date	✅	✅
Review title	✅	— (Google Play has no title field)
App version	✅	✅
Helpful / thumbs-up count	✅	✅
Up-vote sum	✅	—
Developer reply text and date	—	✅
Country and direct review link	✅	✅

Per app — the insights row

• Rating distribution (1★ to 5★) and the average across the reviews collected
• The store-wide rating and rating count, for comparison against your sample
• Per-version breakdown — average rating and volume for each app version, so a bad release is obvious
• Monthly volume — review count and average rating per calendar month
• Negative and positive themes — the words that show up disproportionately in 1-2★ versus 4-5★ reviews
• Developer reply rate — what share of reviews got an answer

Only the new reviews

Turn on onlyNewReviews and the Actor remembers every review ID it has returned and emits only new ones.

Combine it with maxRating: 1 and a schedule and you have a 1-star alarm: every new angry review, within the hour, pushed wherever you want it.

What does it cost?

Pay-per-event. Filters run before charging, so minRating / maxRating / sinceDate genuinely lower the bill.

Event	Price	When
Review scraped	$0.0001	Per review in the dataset
App metadata fetched	$0.002	Once per app, only with metadata on
Insights report	$0.02	Once per app, only with insights on
Actor start	$0.00005	Once per run

Scenario	Cost
One app, both stores, 1,000 reviews each, with insights	~$0.25
10 competitor apps, 500 reviews each, insights on	~$0.72
Daily 1★ watch on 5 apps, ~40 new reviews/day	~$0.02/day

Naming apps

Mix any of these in the apps list:

Form	Example
App Store URL	https://apps.apple.com/us/app/spotify/id324684580
Google Play URL	https://play.google.com/store/apps/details?id=com.spotify.music
iOS numeric ID	324684580
Android package	com.spotify.music
Explicit	ios:324684580 · android:com.spotify.music

Countries matter more than you think

Apple serves at most 500 reviews per country (10 pages of 50) and Google Play's pool is country-scoped too. Countries are the way to scale: ["us","gb","ca","au","de","fr"] gives you six separate pools of the same app.

maxReviewsPerApp is a total across countries, split evenly, so 600 reviews over 6 countries pulls 100 from each.

Output

One dataset, tagged with recordType, with a Reviews view and an Insights view in the Console.

{
  "recordType": "review",
  "platform": "ios",
  "appId": "324684580",
  "appName": "Spotify: Music and Podcasts",
  "country": "us",
  "reviewId": "14407833146",
  "rating": 4,
  "title": "I do love this app but",
  "text": "I do love this app but all I ask is that you please add...",
  "author": "Biker 125",
  "appVersion": "9.1.70",
  "submittedAt": "2026-08-09T19:46:45.000Z",
  "thumbsUp": 0,
  "voteSum": 0,
  "reviewUrl": "https://apps.apple.com/us/app/id324684580?see-all=reviews",
  "scrapedAt": "2026-08-11T01:40:12.004Z"
}

{
  "recordType": "insights",
  "platform": "android",
  "appName": "Spotify: Music and Podcasts",
  "countries": ["us", "gb", "de"],
  "reviewCount": 120,
  "averageRating": 3.63,
  "storeRating": 4.34,
  "storeRatingCount": 36095801,
  "ratingOne": 30, "ratingTwo": 5, "ratingThree": 6, "ratingFour": 10, "ratingFive": 69,
  "negativeThemes": ["paying", "update", "keeps", "anymore", "premium", "listen"],
  "positiveThemes": ["learn"],
  "versionBreakdown": [
    { "version": "9.1.72.1891", "reviewCount": 90, "averageRating": 3.72 },
    { "version": "9.1.68.1888", "reviewCount": 9, "averageRating": 4.11 }
  ],
  "developerReplyRate": 12.5
}

Note the gap between averageRating (3.63 in this run's most-recent reviews) and storeRating (4.34 lifetime). Recent sentiment diverging from the lifetime score is exactly the signal a store rating hides.

Input options

Option	Type	Default	Description
apps	array	required	URLs, IDs, packages or ios: / android: entries
countries	array	["us"]	Two-letter country codes
language	string	""	Google Play language override; Apple follows the country
maxReviewsPerApp	integer	200	Total per app across countries (0 = as many as served)
sortBy	string	recent	recent / helpful / rating (Play only)
minRating / maxRating	integer	1 / 5	Star range to keep
sinceDate	string	""	ISO date, e.g. 2026-01-01
onlyNewReviews	boolean	false	Emit only reviews not seen before
generateInsights	boolean	true	Add the per-app insights row
includeAppMetadata	boolean	true	Look up name, developer, category, store rating
stateStoreName	string	app-reviews-state	Key-value store backing incremental mode
maxConcurrency	integer	5	Parallel requests
proxyConfiguration	object	off	Optional

Use cases

1. Release monitoring — the per-version breakdown shows a regression the day it ships, not when the store rating finally moves.
2. 1-star alerting — maxRating: 1 + onlyNewReviews + a schedule.
3. Competitor teardown — run five rivals at once and read their negative themes to find what their users keep asking for.
4. ASO and localisation — compare ratings and complaints per country before deciding where to invest.
5. Support and product feedback — pipe reviews into your own tooling; the developer reply rate shows where nobody is answering.

Limitations

• Apple caps each country at 500 reviews. Page 11 of its feed returns HTTP 400. Add countries to widen the pool — there is no way around the per-country ceiling.
• Google Play reviews carry no title. The store dropped the field; title is populated for iOS only.
• Developer replies come from Google Play only. Apple's public feed does not include them.
• Theme extraction is rules-based, not an LLM. That keeps it free and predictable, but it counts words rather than understanding them. Function words for English, German, French, Spanish, Portuguese, Italian, Dutch and Turkish are filtered out, along with the app's own name — yet a review pool mixing several languages still produces some noise. For the cleanest themes, run one country at a time. Positive themes are often sparse because 5-star reviews tend to be very short.
• averageRating describes your sample, not the app. It is computed over the reviews this run collected, which with sortBy: recent means recent sentiment. storeRating is the lifetime figure.
• Google Play tolerates datacenter IPs at moderate volume. Apple's feed needs no proxy at all. If you scrape many countries at high frequency, add a proxy for the Play side.
• Apple's feed occasionally repeats reviews across pages. Duplicates are removed by review ID, so a country can return slightly fewer reviews than its page count suggests.