App Store Reviews Analysis: Complaints, Bugs & Bad Updates

Point it at an app and find out what's actually broken — from the people who use it. Give it an app name or App Store id; it pulls the most recent App Store reviews and composes a problem brief: the issue themes users are reporting (crashes, login, update-broke, billing, performance, bugs…), the app version where complaints spiked, the recent rating vs the lifetime rating, and sample complaints. Keyless, agent-callable, built to answer "what are users complaining about in this app right now?"

📊 A signal, not a verdict. It summarizes public App Store reviews (a recent sample) — what users are reporting, not a verified or complete defect list. Review volume and ratings reflect who chose to review.

⚡ Why it's different

A store's lifetime star rating hides current problems. This surfaces them: it contrasts the recent reviews against the all-time average, buckets the complaints into issue themes, and pinpoints the version where things got worse — the answer, not a dump of raw reviews.

📋 What each brief contains

Field	Meaning
signal	recent-regression--investigate / high-recent-complaint-rate / some-recent-complaints / recently-healthy
recentVsOverall	Lifetime rating vs recent average, and the drop
problemReviewRate	% of recent reviews that are ≤2★ or mention a problem
topIssues	Issue themes with counts (crash_freeze, login_auth, update_broke, billing_subscription, performance_slow, bugs_errors, account_data, notifications, ads, ui_ux)
worstVersion	The app version with the lowest recent average rating
byVersion	Per-version review count, average rating, negatives
sampleProblemReviews	Recent low-star reviews with their version and matched issues
overallRating / totalRatings	Lifetime App Store rating and rating count

📥 Input

{ "apps": ["McDonald's", "Bank of America", "324684580"], "country": "us", "maxReviews": 200 }

One brief per app. A name resolves to the closest App Store match; a numeric id is used directly.

📤 Example output (real run, trimmed — McDonald's, US)

{
  "appName": "McDonald's", "overallRating": 4.84, "totalRatings": 7883319,
  "reviewsAnalyzed": 200, "avgRatingRecent": 1.75,
  "signal": "recent-regression--investigate",
  "recentVsOverall": { "overallRating": 4.84, "recentAvgRating": 1.75, "drop": 3.09 },
  "problemReviewRate": 86,
  "topIssues": [ {"issue": "login_auth", "count": 59}, {"issue": "bugs_errors", "count": 57} ],
  "worstVersion": { "version": "26.72.5", "reviews": 200, "avgRating": 1.75, "negativeReviews": 156 },
  "sampleProblemReviews": [ {"rating": 1, "version": "26.72.5", "title": "Update broke my app.",
    "body": "Ever since the new update released I have been unable to sign in to my account...", "issues": ["login_auth","update_broke"]} ]
}

A 4.84-star app whose recent reviews average 1.75 — a sign-in regression the lifetime rating completely hides.

💰 Pricing (pay-per-event)

• App scanned — $0.02 per app for which the brief is produced. Apps that don't resolve are free.

💡 Use cases

• Product / QA / engineering: catch a bad release and the exact version + issue, fast.
• Competitive & market analysts / investors: app health as a live signal for any company's app.
• Agencies / consultants: walk into a pitch knowing exactly what a prospect's app users are complaining about.
• Support / ops: see the complaint themes spiking this week.

🔑 Keyless & agent-callable

No API keys, no accounts, no paid child Actors — just the public Apple App Store search + reviews feeds.

🤖 For AI agents / MCP

https://mcp.apify.com/?tools=actors,docs,centorprat/app-review-problem-scanner

Loads as centorprat--app-review-problem-scanner. Claude Code:

claude mcp add apify -- npx -y @apify/actors-mcp-server --actors centorprat/app-review-problem-scanner

❓ FAQ

Do I need an API key or an Apple account? No. It uses Apple's public App Store search and customer-reviews feeds.

Which reviews are analysed? The most recent ones for the country you pick (Apple serves up to ~500, newest first). It compares them with the app's lifetime rating, so a bad release shows up even when the star rating still looks great.

Google Play? Not yet. This Actor covers the Apple App Store.

Can I compare competitors? Yes. Put several apps in apps; each gets its own brief, so issue themes and recent ratings line up side by side.

Can I monitor an app? Schedule it daily or weekly with a webhookUrl. A recent-regression--investigate signal is your alert that a release went wrong.

Personal data? Reviewer names aren't output. Only ratings, versions, review titles and text are used.

Cost? $0.02 per app with a brief. Apps that don't resolve are free. No paid child Actors.