Apple App Store Reviews & App Feedback Analyzer

Collect bounded public Apple App Store reviews from Apple’s RSS/JSON customer-review feed and turn them into normalized review records, shared-core analysis, platform-scoped clusters, aggregate reports, and cautious release-impact observations.

This Actor is the Apple source prerequisite for the Cross-Platform Mobile App Feedback Intelligence product. It is not an automatic app matcher and it does not scrape authenticated App Store Connect data.

What it supports

• Numeric Apple App Store app IDs and public App Store URLs.
• Storefront country selection and a requested feed locale hint.
• Bounded pagination with review-ID deduplication.
• Raw review records with title, text, rating, date, app version, helpful votes, and source diagnostics.
• Shared feedback-analysis contracts, clustering, per-app reports, and observational release windows.
• Partial source failure: successful apps and already-collected reviews remain available when another request fails.
• English and Vietnamese contract fixtures; the feed locale is preserved as a request dimension.

Input example

{
  "appIds": ["123456789"],
  "country": "US",
  "language": "en",
  "maxReviewsPerApp": 50,
  "maxPagesPerApp": 10,
  "analysis": { "enabled": true },
  "aggregation": { "enabled": true, "minimumClusterSize": 2 }
}

Use appStoreUrls when the numeric ID is not already available. Explicit IDs take precedence over duplicate URL IDs.

Release impact

Set mode to releaseImpact and provide release.releasedAt, daysBefore, and daysAfter. The result compares review windows around the release timestamp and uses observational language such as “issue mentions increased after release.” It does not prove causation.

Output

Dataset records include review, sourceDiagnostic, feedbackCluster, productFeedbackReport, and feedbackImpactReport records. Normalized records retain the source platform as apple-app-store, the Apple app ID, original text, source locale, storefront country, and nullable fields where Apple does not expose metadata.

RUN_STATS records collection, analysis, aggregation, runtime, memory, and error counts. Per-app reports are stored under APP_STORE_REPORT_<app-id>.

Limitations and responsible use

• The public feed is bounded and storefront-specific; it is not a complete worldwide review archive.
• Apple’s public feed does not guarantee the reviewer’s original language. language is retained as the requested feed locale and must not be treated as ground-truth reviewer language when it is unavailable.
• Developer replies are not exposed by this public feed and remain null unless a future permitted source provides them.
• Missing app-version, country, language, or date fields remain null/unknown; the Actor does not invent metadata.
• Detected issues are user reports, not confirmed defects. Important product decisions require manual validation.
• AI or fallback analysis may misclassify sarcasm, ambiguity, or mixed-language feedback.

Local development

npm install
node --test test/*.test.mjs
apify validate-schema

The Actor uses Apple’s public RSS/JSON customer-review feed and does not require an API key for public review collection. Respect Apple’s current feed terms, rate limits, and applicable laws.