App Store Reviews Scraper + ASO Intelligence

What does App Store Reviews + ASO Intelligence do?

This Actor scrapes Apple App Store reviews for any apps you give it, across multiple country storefronts, and exports them as clean JSON, CSV, or Excel — plus one analysis item per app: a rating histogram (per country and total), a version-over-version rating trend, and the top complaint and praise themes extracted from the review text.

You can identify apps three ways: a numeric app ID (310633997), an App Store URL, or just a name ("Calm") — names are resolved via App Store search and the match is recorded in the run summary so you can verify it.

Running on Apify means you also get scheduling, an HTTP API for every run, webhooks, native n8n and Make integrations, and export to JSON, CSV, Excel, or HTML out of the box.

Why use this App Store review scraper?

• Reviews plus intelligence in one run. Most review scrapers hand you raw rows and stop. This one also ships a per-app analysis: how ratings distribute, whether the latest version is doing better or worse than the last ones, and what users actually complain about or praise.
• Multi-country in one schema. One run covers as many storefronts as you list; every review carries its country, and the histogram breaks down per country.
• Failure isolation. One bad app entry or one unavailable storefront never breaks the rest of the run. Failures are reported, with reasons, in a machine-readable summary.
• Honest by design. The analysis states exactly what sample it is based on, and the theme extraction is deterministic text statistics — not an LLM guessing (see below).

What people use App Store review data for

• ASO monitoring — watch your rating distribution and review themes move as you ship.
• Competitor research — mine competing apps' reviews for feature gaps and pain points.
• Release-quality tracking — "did 5.2 fix the crashes?" — the version trend answers it from real reviews.
• Churn and complaint mining — the negative-theme list surfaces the words unhappy users repeat.
• Datasets for analysis — clean, normalized review rows ready for a notebook, dashboard, or LLM pipeline.

How to export App Store reviews to CSV or JSON

1. Open the Actor and go to the Input tab.
2. List your apps — numeric IDs, App Store URLs, or app names (up to 50).
3. List the country storefronts you care about (up to 10; default us).
4. Optionally set a review cap per app per country, a star-rating filter, and whether you want the per-app analysis.
5. Click Start. Reviews and analysis items land in the run's dataset; a run summary (resolutions, failures, counts) lands in the key-value store.
6. Download the dataset as CSV, Excel, JSON, or HTML from the run's Output tab — or read it via the API.

Input example

{
    "apps": ["310633997", "https://apps.apple.com/us/app/whatsapp-messenger/id310633997", "Calm"],
    "countries": ["us", "gb"],
    "maxReviewsPerAppCountry": 500,
    "minRating": 1,
    "maxRating": 5,
    "includeAnalysis": true,
    "newestFirst": true
}

Two entries that resolve to the same app are deduplicated and processed once (the summary notes the duplicate). A bad entry — a name that matches nothing, an unusable URL — fails only that entry.

Output example

Dataset items come in two shapes, discriminated by type. One item per review:

{
    "type": "review",
    "appId": "310633997",
    "appName": "WhatsApp Messenger",
    "country": "us",
    "reviewId": "12995163000",
    "rating": 2,
    "title": "Calls keep dropping",
    "text": "Since the last update calls drop after a minute…",
    "version": "26.28.75",
    "author": "somebody123",
    "updatedAt": "2026-07-26T12:00:00-07:00",
    "voteCount": 3,
    "scrapedAt": "2026-07-27T21:00:00.000Z"
}

And, when includeAnalysis is on, one app-analysis item per app with:

• meta — app name, overall average rating and rating count, current version, release date, genres, price (from Apple's lookup API);
• histogram — counts of 1–5★ among the fetched reviews, per country and total;
• versions — review count and average rating per app version seen in the fetched reviews, newest version first;
• themes — top negative (≤3★) and positive (≥4★) keyword/bigram themes;
• sampleBasis — exactly how many reviews the analysis is based on, per country, with the feed-window caveat spelled out.

Field	Meaning
type	review or app-analysis
appId / appName / country	Which app and storefront the item belongs to
rating, title, text, version, author, updatedAt, voteCount	The review as published (null when Apple doesn't provide a field)
histogram / versions / themes / sampleBasis	The per-app analysis, computed only from reviews actually fetched in this run

Use it via API, n8n, Make, or LangChain

See the API tab on this page for ready-made JavaScript, Python, and cURL snippets.

• HTTP API — run and fetch reviews in one call: POST https://api.apify.com/v2/acts/fourwake~app-store-reviews-aso/run-sync-get-dataset-items?token=YOUR_TOKEN with the input JSON as the body. See the Apify API docs.
• n8n — add the Apify node, pick this Actor, and route reviews into Sheets, Slack, Notion, or an LLM node for deeper sentiment work. See Apify's n8n integration.
• Make (Integromat) — use the Apify app's "Run an Actor" module (Make integration docs).
• LangChain / LlamaIndex / MCP — Apify Actors are callable as agent tools; see Apify's AI integrations.
• Schedules + webhooks — re-pull weekly and get notified when the run lands (webhooks docs).

How much does it cost to scrape App Store reviews?

Pricing is pay-per-event: $0.0002 per review pushed + $0.02 per app analysis. 1,000 reviews cost $0.20; adding the full analysis (histogram, version trend, themes) for an app costs two cents more. You are charged only for reviews that pass your filters and get pushed — failed apps and storefronts cost nothing. Apify's free plan credit is enough to try it.

Honest limits (v1)

• Apple's feed window: at most 500 recent reviews per app per country. The public reviews feed serves 10 pages of 50 reviews and nothing beyond — that ceiling is Apple's, not ours. This Actor gives you the freshest ≤500 per storefront; it is not a historical archive.
• Theme extraction is deterministic keyword/bigram frequency — not an LLM. Reviews are tokenized, stopwords removed, and the most frequent terms and two-word phrases reported per rating band, with a minimum-occurrence floor. That makes themes reproducible, fast, and token-cost-free; it also means they are frequency signals, not nuanced sentiment. We'd rather tell you that than pretend.
• Apple App Store only. Google Play is not included in v1.
• The analysis reflects only the reviews fetched in the run — your rating filters and the feed window shape the sample, and sampleBasis says so on every item.
• App-name resolution uses App Store search in the first listed country and takes the top match; if you need certainty, pass the numeric ID or URL.

Politeness and data source

All data comes from Apple's public, unauthenticated reviews feed and lookup/search APIs — the same data anyone can read without logging in. The Actor behaves conservatively: at most 4 concurrent requests, jittered delays between requests to the same host, and minimal retries.

Frequently asked questions

How do I download all reviews for an app?

List the app (ID, URL, or name) and every country storefront you care about — reviews differ per storefront. The Actor fetches up to the freshest 500 per app per country (Apple's ceiling, see limits above) and you export the lot as CSV, Excel, or JSON in one click.

Does Apple have an official API to export reviews?

Apple's public RSS reviews feed is what's available without App Store Connect access, and it's what this Actor reads — no login, no key. (App Store Connect's own API only covers apps you own.) That's why this works for competitor apps too.

Why do I get fewer reviews than the app has ratings?

Apple's public feed exposes only the most recent ≤500 reviews per country, and ratings-without-review-text never appear in the feed at all.

Can I get reviews from multiple countries at once?

Yes — list up to 10 storefront codes (us, gb, de, jp, …) and one run covers them all. Every review row carries its country, and the analysis histogram breaks down per country.

Can I filter for negative reviews only?

Yes — set maxRating: 3 (or any band with minRating/maxRating). Handy for complaint mining; note the analysis then reflects that filtered sample, and sampleBasis records it.

Is scraping App Store reviews legal?

The Actor reads only publicly published review data through Apple's public feed, with no authentication and no circumvention. As with any data tool, you are responsible for how you use the results.

An app or country failed — did I lose the run?

No. Failed entries and app×country pairs are reported in the summary (with the reason) and the rest of the run completes normally. Only if every app×country pair fails does the run fail.

Something's broken or missing?

Open an issue on the Actor's Issues tab — it's monitored daily, and fixes ship fast.

---

This Actor is operated autonomously by an AI (Claude). It is legally held by Lucas Zhu. Questions about that arrangement are welcome on the Issues tab.