Apple App Store Review & Competitor Intelligence

Monitor recent public Apple App Store written reviews across apps and storefronts. The Actor reads Apple's public customer-review RSS/JSON feed, normalizes reviews, and adds lightweight deterministic intelligence for sentiment, complaints, competitor mentions, switching signals, and newly observed reviews.

Why this Actor

• No Apple developer account or App Store Connect token required for the public review-feed path.
• Direct HTTP first: no browser and no proxy cost when Apple's public feed works normally.
• Cost-bounded memory profile: 256 MB default, 256 MB minimum, 512 MB maximum for this HTTP-only workload.
• Bounded Apify/custom proxy fallback for confirmed blocks or transient network failures.
• Multi-app and multi-country storefront monitoring.
• Recent review intelligence: rating, title, body, app version, date, sentiment, complaint categories, competitor mentions, switching signals.
• Optional reviewer display name; off by default.
• Monitoring state for newly delivered reviews with a bounded cumulative ID ledger, preventing repeat billing if Apple reorders review windows.
• Cache for non-monitoring runs.

Important source limitation

Apple's public RSS customer-review feed is an undocumented public surface. It exposes a recent window per storefront rather than a complete historical archive. In practice it is commonly paginated in roughly 50-review pages with a limited recent window. The Actor therefore describes its results as recent public reviews, not complete lifetime history.

Input example

{
  "apps": ["389801252"],
  "countries": ["us", "gb"],
  "maxReviewsPerApp": 50,
  "sort": "mostrecent",
  "competitorNames": ["TikTok", "Snapchat"],
  "monitorChanges": true
}

Apps may be numeric IDs or App Store URLs containing id<digits>.

Output

review

One row per successfully returned review. Complaint intelligence includes stability/performance, connectivity, ads, pricing, login/account, usability, content moderation, and support themes when supported by review text. Common explicit non-complaint phrases such as “no crashes” or “never had a problem” are masked to reduce false positives. Complaint detection is independent from star-led sentiment, so a positive-rated review can still expose a concrete complaint. Includes appId, country, reviewId, title, reviewText, rating, appVersion, publishedAt, sentiment, complaints, switching data, and competitor mentions.

app_summary

One summary per app/storefront pair with recent average rating, sentiment counts, complaint-category counts, competitor mentions, and new-review count when monitoring has a baseline.

error

Structured diagnostics for invalid inputs, exhausted request budgets, blocks, transient upstream failures, or network errors. Error and summary rows are not charged by the review event.

Reliability design

• Actor-scoped cache and monitoring state for Limited Permissions.
• PPE-aware delivery: charged batches honor the caller spending limit, stop before extra network work when exhausted, and never mark unpushed reviews as delivered.
• Hard request budget.
• Direct-first Apple feed route.
• Bounded retry/backoff and fresh proxy sessions for 401/403/429, HTTP 408/5xx/520-524, and network timeouts.
• JSON feed primary path with Atom XML format fallback.
• Empty-page tolerance so one empty page does not automatically terminate pagination.
• No CAPTCHA solving.

Pricing

Suggested pay-per-event price: $0.0015 per successfully returned review intelligence row. App summaries, errors, retries, empty pages, proxy rotations, and monitoring state updates are not charged by that event.

Privacy

Only public App Store review data is processed. Reviewer display names are omitted by default and can be enabled explicitly.

Disclaimer

Independent community Actor. Not affiliated with, endorsed by, or operated by Apple Inc.