App Store Reviews Scraper + AI Analysis

Scrape Apple App Store reviews for any app in any country (no login) and get optional AI analyses on top: pain points, feature requests, rating trend, and a headline summary you can paste into a report.

Most App Store scrapers hand you one country and a pile of raw rows. This one sweeps every storefront in a single run and ships the analysis you were going to build anyway.

---

Why more countries means more data

Apple caps its public review feed at 200 reviews per app, per storefront. That cap is per storefront — and every storefront holds a different set of reviews.

So the same app that tops out at 200 reviews in us gives you 1,667 once you sweep ten countries. Ask for ["all"] and you get every storefront Apple serves.

Countries requested	Reviews returned (Notion, measured on Apify)
us only	200 (Apple's per-storefront cap)
10 storefronts	1,667

Some storefronts return nothing for a given app — the actor skips them and keeps going. Ten requested, nine returned data, in 61 seconds.

Version regressions — the part nobody else does

Every release, someone asks "did that update hurt us?" This actor answers it directly:

{
  "regressions": [
    { "fromVersion": "1.7.233", "toVersion": "1.7.234",
      "before": 4.1, "after": 2.625, "ratingDelta": -1.475,
      "sampleBefore": 10, "sampleAfter": 8 }
  ],
  "worstVersion": { "version": "1.7.271", "averageRating": 1.714 }
}

Per-version stats include a Wilson lower bound, so a version with 5 reviews doesn't outrank one with 500.

Analyses

Each is billed once per run, not per review.

Analysis	What you get
versionRegressions	Per-version ratings, ranked drops, worst/best release
painPoints	What ≤3-star reviewers actually complain about, scored by how much more often a term shows up in negative reviews than overall (lift), with a real example quote
featureRequests	What users ask for, extracted from reviews containing request language
summary	Headline numbers: volume, average, negative share, biggest regression

Analysis is deterministic — no LLM in the loop. Same input, same output, every time. No hallucinated themes, no per-token cost passed on to you.

Input

{
  "appIds": ["1232780281"],
  "countries": ["all"],
  "sortBy": "mostRecent",
  "maxReviewsPerCountry": 200,
  "minRating": 1,
  "maxRating": 3,
  "sinceDate": "2026-01-01",
  "aiAnalysis": ["versionRegressions", "painPoints"]
}

appIds accepts numeric IDs or full App Store URLs. No IDs handy? Pass searchTerm instead and the actor resolves it.

Output

Reviews (type: "review"):

Field
appId appName country	which app, which storefront
rating title body version	the review
author updatedAt voteSum voteCount url	metadata

Analyses arrive as separate items with type: "analysis".

Pricing

Event	Price
Actor start	$0.00005
Per review	$0.0002 ($0.20 per 1,000)
Per analysis	$0.05 (summary: $0.03)

A 5,000-review sweep with two analyses runs about $1.10. A single-country pull with one analysis is about $0.09.

Why this one won't rot

It reads Apple's own public RSS feed — the same endpoint Apple publishes for anyone to use. No login, no headless browser, no proxy rotation, no CAPTCHA arms race. Nothing to break when a site ships a redesign.

It also means no personal data: App Store review authors are public display names, and this actor collects nothing beyond what Apple already publishes.

Notes

• Some storefronts return zero reviews for a given app. That's normal — the actor skips them and continues.
• Google Play is deliberately not supported: Google's review API only covers apps you own, and scraping it is a terms violation. This actor stays on the clean side of that line.

---

Built by M＆Y.