App Store Reviews & Charts Monitor

Monitor any Apple App Store app across any App Store countries. You get recent customer reviews (star rating, title, text, app version, date), app details and rating, and top-chart ranks. Give it app ids, App Store links, bundle ids or plain app names.

It is built for monitoring. On a schedule it can return only the reviews that are new since your last run. If a run stops early, the next run picks up the reviews it skipped. Each app record also includes a rating breakdown per app version, so you can see whether the latest release made ratings worse. Chart ranks follow the App Store's own Apps charts: Top Free and Top Paid exclude games, as the App Store does. Every record is flat JSON with a fixed set of fields, so AI agents, Zapier, Make, n8n or a spreadsheet can use it directly.

It uses only Apple's public, no-login endpoints: the iTunes Search/Lookup API and Apple's RSS chart and review feeds. There are no browsers, no scraped tokens and no accounts.

Who it's for:

• App developers and product teams watching their own reviews and competitors'.
• ASO and marketing people tracking chart ranks, screenshots and listing changes.
• Support teams that want 1–2★ reviews sent to Slack or a helpdesk.
• AI agents that need current App Store data.

What you get

Each dataset row has a record_type:

record_type	One row per	Charged event
app	app × storefront: details, rating, and an aggregate of the reviews scanned	app-record
review	customer review	review
chart	chart position, or one row per tracked app with chartTrackedAppsOnly	chart-row
notice	explanation of something skipped: bad id, unknown country, empty feed…	free

All ids (app_id, developer_id, review_id, genre_id, category_id) are strings.

Review fields

• app_id, app_name, country, review_id
• rating (1–5), title, text
• author: the reviewer's public App Store nickname
• app_version, date (UTC ISO), vote_sum, vote_count, fetched_at
• is_new: true when no earlier monitoring run with the same stateKey returned this review (see the FAQ)

App fields

• Identity: app_id, country, name, developer, developer_id, seller_name, bundle_id.
• Price and category: price, currency, formatted_price, primary_genre, genres, content_rating, advisories. Genre names are in English in every storefront. A genre Apple adds later keeps Apple's own label until the actor knows it.
• Versions: version, current_version_release_date, release_date.
• Rating: rating_avg and rating_count (lifetime, star-only ratings included). rating_avg is null when the app has no ratings. Both are null for Mac apps, because Apple's Lookup API does not report their ratings. Use reviews_avg_rating for Mac apps.
• Size and compatibility: file_size_mb (decimal MB, as the App Store shows it), minimum_os_version, languages.
• Links and media: app_url, icon_url, screenshot_urls, ipad_screenshot_urls (up to 10 each), developer_url.
• Text: release_notes, description, both complete (the App Store's limit is 4,000 characters).
• matched_search_term: the search term that found the app in this storefront. It is null for apps given as ids, URLs or bundle ids.

When reviews are fetched in the same run, the record also has:

• reviews_scanned, reviews_new, reviews_avg_rating, reviews_histogram
• reviews_oldest, reviews_newest
• version_ratings: reviews and average stars per app version, newest version first

These cover every review the run scanned for that app and storefront, from the newest down to where it stopped. The star filter and onlyNewReviews then choose which of them are returned as rows. So with ratings: ["negative"] you get only 1–2★ rows, but reviews_scanned, the histogram and version_ratings still describe all the reviews scanned. The fields are null when reviews are not fetched.

Chart fields

• chart, country.
• genre_id and genre_name: what the chart ranks.
  • All apps (games excluded) for overall free and paid charts.
  • All apps and games for grossing.
  • All Mac apps for Mac.
  • The genre name for genre charts.
• rank: the position in Apple's chart feed (see the FAQ on app bundles). With chartTrackedAppsOnly, null means the app is not in the first chart_depth positions.
• chart_depth: how many positions were scanned.
• app_id, app_name, developer, bundle_id, category (the app's genre in English) and category_id, price, currency, release_date, app_url, icon_url, is_tracked_app.
• source: which Apple feed answered.
• fetched_at.

Real output (runs of 2026-09-25)

App record (app name pinterest, US storefront, 500 reviews fetched; the screenshot and language lists are shortened here):

{
  "record_type": "app",
  "app_id": "429047995",
  "country": "us",
  "name": "Pinterest",
  "developer": "Pinterest, Inc.",
  "developer_id": "328135727",
  "seller_name": "Pinterest, Inc.",
  "bundle_id": "pinterest",
  "price": 0.0,
  "currency": "USD",
  "formatted_price": "Free",
  "primary_genre": "Lifestyle",
  "genres": ["Lifestyle", "Photo & Video"],
  "content_rating": "12+",
  "advisories": ["Infrequent/Mild Medical/Treatment Information", "Infrequent/Mild Cartoon or Fantasy Violence", "Infrequent/Mild Sexual Content and Nudity", "Infrequent/Mild Alcohol, Tobacco, or Drug Use or References", "Infrequent/Mild Profanity or Crude Humor", "Infrequent/Mild Horror/Fear Themes", "Infrequent/Mild Mature/Suggestive Themes"],
  "version": "14.36",
  "current_version_release_date": "2026-09-16T15:11:28Z",
  "release_date": "2011-04-28T15:51:13Z",
  "rating_avg": 4.778,
  "rating_count": 5776573,
  "file_size_mb": 425.3,
  "minimum_os_version": "16.0",
  "languages": ["AF", "AR", "BN", "BG", "HR", "CS", "DA", "NL", "EN", "FI", "FR", "DE", "…"],
  "app_url": "https://apps.apple.com/us/app/pinterest/id429047995",
  "icon_url": "https://is1-ssl.mzstatic.com/image/thumb/Purple211/v4/7d/d9/a2/7dd9a2dc-88a8-9ced-4a33-f06f9d0091bd/AppIcon-0-0-1x_U007epad-0-1-0-0-0-85-220.png/512x512bb.jpg",
  "screenshot_urls": ["https://is1-ssl.mzstatic.com/image/thumb/PurpleSource211/v4/07/43/95/074395db-1ab5-4adf-ec05-037f91216984/iPhone-5.5_1.png/392x696bb.png", "…"],
  "ipad_screenshot_urls": ["https://is1-ssl.mzstatic.com/image/thumb/PurpleSource211/v4/2e/d2/90/2ed29067-5da9-5761-5977-9ef54cfb4955/iPad-12.9_1.png/576x768bb.png", "…"],
  "developer_url": "https://apps.apple.com/us/developer/pinterest-inc/id328135727",
  "release_notes": "Every week, we polish up the Pinterest app. This update includes:\n- Bug fixes\n- Performance improvements\n\nTell us if you like this latest version at https://help.pinterest.com/contact",
  "description": "Pinterest is a place of endless possibilities. You can:\n- Discover everyday inspiration\n- Shop styles you love\n- Try and learn something new\n\nCreate boards, save Pins and make collages of all your inspiration. Unlock billions of ideas, from fashion tips and easy recipes to DIY projects and fresh ways to redo your space. Creating the life you love? \n\nIt's Possible.",
  "matched_search_term": "pinterest",
  "reviews_scanned": 500,
  "reviews_new": 500,
  "reviews_avg_rating": 2.36,
  "reviews_histogram": {"1": 231, "2": 84, "3": 53, "4": 37, "5": 95},
  "reviews_oldest": "2026-09-11T22:35:07Z",
  "reviews_newest": "2026-09-23T23:02:55Z",
  "version_ratings": [
    {"version": "14.36", "reviews": 297, "avg_rating": 2.47},
    {"version": "14.35", "reviews": 203, "avg_rating": 2.2}
  ],
  "fetched_at": "2026-09-25T02:15:11Z"
}

In this example the lifetime rating is 4.78★, but the last 500 written reviews average 2.36★. The aggregate fields exist to surface gaps like this.

Review:

{
  "record_type": "review",
  "app_id": "429047995",
  "app_name": "Pinterest",
  "country": "us",
  "review_id": "14585540524",
  "rating": 1,
  "title": "fix your app!!!!",
  "text": "its good but why cant i see my pins from my main account on my alt account?? fix your app people cant even see my pins! your updating anything but your own app",
  "author": "maxine 🌸",
  "app_version": "14.36",
  "date": "2026-09-23T22:00:05Z",
  "vote_sum": 0,
  "vote_count": 0,
  "is_new": true,
  "fetched_at": "2026-09-25T02:15:11Z"
}

Chart row. Rank tracking for RadarScope (288419283) in US Top Paid Apps:

{
  "record_type": "chart",
  "chart": "top-paid",
  "country": "us",
  "genre_id": null,
  "genre_name": "All apps (games excluded)",
  "rank": 5,
  "chart_depth": 100,
  "app_id": "288419283",
  "app_name": "RadarScope",
  "developer": "Base Velocity, LLC",
  "bundle_id": "com.basevelocity.RadarScope",
  "category": "Weather",
  "category_id": "6001",
  "price": 9.99,
  "currency": "USD",
  "release_date": "2019-02-12",
  "app_url": "https://apps.apple.com/us/app/radarscope/id288419283",
  "icon_url": "https://is1-ssl.mzstatic.com/image/thumb/Purple221/v4/df/7f/31/df7f31db-6974-a1fb-748d-f31b4d3c8f7e/App_Icon-0-0-1x_U007emarketing-0-0-0-8-0-0-85-220.png/100x100bb.png",
  "is_tracked_app": true,
  "source": "marketingtools-rss",
  "fetched_at": "2026-09-25T02:15:23Z"
}

The dataset has four table views: Reviews, Apps, Charts and Notices. You can export as JSON, CSV or Excel.

Input

Field	Type	Default	What it does
apps	list	ChatGPT (6448311069), only when apps, searchTerms and every chart option are empty	App Store ids, App Store URLs, bundle ids (com.burbn.instagram) or app names. Ids, URLs and bundle ids name one exact app. App names are searched for in each storefront (see the FAQ). Up to 100.
searchTerms	list	—	Names or keywords to search for. Each is searched in every storefront, and that storefront's top result(s) are monitored there. Up to 20.
countries	list	us	Storefront codes (us, gb, de, jp…) or country names. With no countries set, the country in an App Store URL is used. Max 50.
modes	list	see below	Any of metadata, reviews, charts.
maxReviewsPerApp	int	100	Maximum reviews per app per storefront, newest first. Max 500, which is Apple's limit. On a monitor's first run it also sets how far back the monitor starts.
onlyNewReviews	bool	false	Return, and charge for, only reviews that earlier monitoring runs have not returned.
ratings	list	all	Star filter, e.g. ["1","2"], ["negative"] (1–2), ["positive"] (4–5), ["1-3"].
chartTypes	list	top-free	top-free, top-paid, top-grossing, top-free-ipad, top-paid-ipad, top-grossing-ipad, top-free-mac. A name that mentions games, such as top free games, gives that chart in the Games genre.
chartGenres	list	overall	Genre names or ids: Games, Productivity, Finance, Puzzle games, 6014…
chartLimit	int	25	Rows per chart (1–100).
chartTrackedAppsOnly	bool	false	Rank tracking. Returns one row per app, chart and storefront instead of the whole chart.
searchResultsPerTerm	int	1	How many search results to monitor per term and storefront (max 10).
maxRecords	int	—	Hard cap on charged records per run.
stateKey	string	default	Name of a monitor. It keeps separate "new review" state for separate monitors of the same app.

Which modes run. Without modes, the actor fetches metadata and reviews. If you set any chart option (chartTypes, chartGenres or chartTrackedAppsOnly) and leave modes empty, it fetches only charts. The apps you list then only mark their chart rows (is_tracked_app), or are the apps ranked with chartTrackedAppsOnly. They are not charged as app records. If you set modes and a chart option, charts are added to your modes, and a free notice row says so. In the Console form, Apps is prefilled with ChatGPT and What to fetch is empty, so adding a chart there fetches charts only.

Input is lenient:

• Values are case-insensitive and trimmed.
• A list entry can hold several comma-separated values ("us, gb").
• Anything the actor cannot use is skipped and explained in a free notice row, and the run does not fail because of it. Up to 20 such notices are listed per field; the rest are counted in one more notice.
• Unusable apps, search terms, chart types and genres are never swapped for something else. If you name apps, search terms or charts and none of them can be used (a URL without an id, a Google Play link, an unknown chart or genre), the run returns only notice rows and charges nothing. The example app ChatGPT is used only when the input names no apps, search terms or chart options at all. Two settings do fall back, with a notice: an unusable countries list becomes us (or the storefront in an App Store URL), and unusable modes become the default.
• App names and search terms return whatever App Store search returns in each storefront. The FAQ explains how the actor guards against Apple's stand-in results.

Apify checks the fields above against the input schema before the actor starts. A list field given as a single string is rejected: "ratings": "negative" fails, so send ["negative"]. Apify does not check unknown keys, so the actor's aliases work with either a list or a single string: appIds, country, maxReviews, onlyNew, stars, genre and more.

Examples:

{ "apps": ["https://apps.apple.com/us/app/duolingo-language-lessons/id570060128", "com.burbn.instagram", "Notion"],
  "countries": ["us", "gb", "de"], "maxReviewsPerApp": 500 }

{ "apps": ["6448311069", "6473753684"], "searchTerms": ["google gemini"], "modes": ["reviews"],
  "ratings": ["negative"], "onlyNewReviews": true, "stateKey": "slack-bad-reviews", "countries": ["us", "ca", "au"] }

{ "apps": ["Notion", "Todoist"], "chartGenres": ["Productivity"], "chartTrackedAppsOnly": true, "countries": ["us", "gb"] }

{ "chartTypes": ["top-free", "top-grossing"], "chartGenres": ["all", "Games"], "countries": ["us", "jp"], "chartLimit": 100 }

Pricing

Pay per event. There is no subscription or rental fee, and platform usage is included.

Event	Price	When
actor-start	$0.005	once per run, charged with the run's first charged record
app-record	$0.002	per app per storefront (metadata mode)
review	$0.0004	per review returned
chart-row	$0.0005	per chart row returned
notice rows	free	skipped input and upstream problems

Worked examples:

Run	Records	Cost
Default (ChatGPT, US, 100 reviews)	1 app + 100 reviews	$0.005 + $0.002 + $0.04 = $0.047
Deep pull: 3 apps × 3 countries × 500 reviews	9 apps + 4,500 reviews	$0.005 + $0.018 + $1.80 = $1.82
Daily monitor: 5 apps × 3 countries, onlyNewReviews, ~20 new reviews each	15 apps + ~300 reviews	$0.005 + $0.03 + $0.12 ≈ $0.16/day
Top 100 free + grossing charts in US + GB	up to 400 chart rows	$0.005 + $0.20 = $0.205
Rank tracking: 3 apps, 2 charts, 3 countries (no modes)	18 chart rows	$0.005 + $0.009 = $0.014

A run that returns no charged records costs nothing, not even the start fee. That covers a monitor run with no new reviews (in reviews-only mode), app ids Apple does not know, and input where nothing could be used.

You can set a maximum charge per run in the run options, or use maxRecords. The actor stops cleanly at that limit. It keeps everything it has already fetched and ends as SUCCEEDED. If the maximum charge cannot cover the $0.005 start fee plus one record ($0.0054 for a review), the actor fetches nothing and charges nothing.

FAQ

How many reviews can I get?

Apple's public review feed returns the 500 most recent written reviews per app per country, as 10 pages of 50. To go further back, add more countries. To keep a complete history from now on, schedule the actor with onlyNewReviews and add each run's rows to your own table. Very busy apps (ChatGPT, Instagram) collect 500 reviews in 1–5 days, so reviews beyond Apple's 500 are out of reach even with a daily schedule.

How does "new since last run" work?

A monitoring run is a run with onlyNewReviews: true or with a stateKey. Monitoring runs save the ids of the reviews they returned in a named key-value store in your Apify account, called app-store-reviews-monitor-state. The ids are stored per app, storefront and stateKey. Other runs only read this state, so an ad-hoc run never uses up a scheduled monitor's new reviews.

Where a monitor starts. The first monitoring run for an app and storefront looks at the newest maxReviewsPerApp reviews that pass the star filter. That window sets where the monitor starts: it never returns reviews older than the oldest review in the window. To start further back, give the first run a larger maxReviewsPerApp (up to 500).

A review is new (is_new: true) when no earlier monitoring run returned it and it is not older than where the monitor starts. On a monitor's first run every review is new. Apple sometimes publishes a review a day or more after its timestamp, so the check allows 3 days for that.

Runs that stop early. With onlyNewReviews: true, only new reviews are returned and charged. Paging stops once the feed reaches reviews the monitor already covered. If a run stops before it has returned everything, the next run returns the reviews it skipped, as long as Apple still lists them among its 500 most recent. This includes the first run: if your maximum charge stops it after 50 of its 500 reviews, the next run returns the other 450. A run can stop early because of maxRecords, your maximum charge, a timeout or an Apple error. After the first run, maxReviewsPerApp can stop it early too. If a timeout or an Apple error stops the first run before it has looked at its whole window, the next run still counts as the first.

Reviews removed by the star filter count as handled. Give each differently filtered monitor its own stateKey.

How are app names, search terms and bundle ids matched?

• An App Store id or URL names one app. It is fetched in every storefront you list, or only in the URL's storefront when you set no countries. Where the app is not sold, you get a free notice.
• A bundle id is looked up in your first storefront and, if Apple does not find it there, in the next ones. The app it names is then fetched like an id.
• An app name or search term is searched in each storefront separately, and that storefront's top result is fetched there. matched_search_term shows which term found it. Apple's search is fuzzy and app names differ by country, so the results can differ between storefronts. The actor applies two checks:
  • Sometimes a storefront does not sell the app, and its top result is some other app. For example, searching Venmo in gb returns PayPal, while us returns Venmo. When the term appears in the top result's name in other storefronts but not in this one, nothing is fetched there, and a free notice explains why.
  • When the storefronts' top results are different apps that all contain the term, each storefront keeps its own result, and a free info notice lists them.
• With only one storefront, or when no storefront's top result contains the term, there is nothing to compare, so Apple's top result is used. With searchResultsPerTerm above 1, the checks are off and every storefront keeps all its results. Use ids or URLs when you need one exact app.
• Search is slow. Apple allows about 20 searches a minute, so each term takes about 3 s per storefront. After 200 searches (for example 20 terms × 10 storefronts), the remaining storefronts reuse the app each term found, and a free notice says so.

How fresh are the reviews? How often should I schedule?

Apple's review feed behaves like a daily snapshot. In our tests (September 2026), the newest review in the feed was about 1–1.5 days old. Every app we tried stopped at the same cutoff, and the feed was identical two hours apart. Scheduling more often than once a day therefore returns nothing new. The actor returns what the feed contains.

Are star-only ratings included?

No. The feed contains written reviews only. The app record's rating_avg and rating_count include star-only ratings. They are lifetime values: Apple's public Lookup API no longer gives a separate current-version rating. For the rating of each version, use version_ratings, which is built from the written reviews scanned.

What is version_ratings?

It groups the reviews scanned in this run by the app version the reviewer used and averages their stars. You can use it to see whether the latest release changed sentiment.

Do the chart ranks match the App Store?

Almost. The App Store has separate Apps and Games charts:

• top-free, top-paid, top-free-ipad and top-paid-ipad without a genre are the Apps charts, so games are excluded. We checked them against the App Store charts pages on 2026-09-24, and the US top 25 matched position for position on iPhone and iPad.
• For game charts, use the genre Games (or a game genre such as Puzzle games).
• top-grossing has no App Store equivalent any more. It is Apple's combined ranking of apps and games, labelled All apps and games.

App bundles are never in Apple's chart feeds. A bundle is several apps sold together. The App Store charts show bundles, but Apple's feeds leave them out, so an app ranked below a bundle shows one place higher than on the App Store. This mostly affects paid charts. For example, in September 2026 a driving-theory test bundle sat around #20 of UK Top Paid Apps, so the actor's ranks from there down were one place higher. The actor returns Apple's feeds as they are and cannot see where the bundles rank.

top-paid comes from Apple's newer chart feed (source: marketingtools-rss), which covers about 100 positions. That feed is slow on its first request for a storefront (up to about 30 s) and sometimes times out. The actor retries it, and if it still does not answer, uses Apple's older feed with games removed (source: itunes-rss) and adds a free notice row. The older feed reaches only about 20–60 positions deep for paid apps, so with chartTrackedAppsOnly an app ranked below that gets rank: null in that run. If the newer feed fails for 2 storefronts in a row, the actor stops asking it for the rest of the run. The iPad Apps charts have the same limit (about 30–70 positions). chart_depth tells you how deep each chart went.

Rows from the newer feed do not carry price, currency, bundle id or category. The actor fills them from the older feed, or from Apple's Lookup API for apps outside the older feed's top 100, so these fields are filled on every ranked row unless that lookup fails.

Why is rank null?

With chartTrackedAppsOnly, a row with rank: null means that the app is not in the first chart_depth positions of that chart in that storefront. Such a row carries the app's id and name only. The other app fields (price, bundle_id, category and so on) are null there. Add metadata to modes if you also want the app details.

Which countries work?

Any App Store storefront, given as a 2-letter code (us, gb, de, jp, br, in…). Common country names such as "United Kingdom" or "Germany" also work. If Apple rejects a code, you get a free notice row, and the rest of the run continues.

Does it cover Google Play?

No. It covers the Apple App Store only.

Do I need an API key or an Apple account?

No. It uses Apple's public iTunes Search/Lookup API and public RSS feeds. It sends an honest User-Agent and paces its requests:

• Search/Lookup: about one call every 3 s, since Apple documents about 20 calls/min.
• Feeds: about 3 calls per second.

Apple's review feed sometimes returns an empty page that its CDN then caches for a few minutes. The actor detects these and asks for the same page again under Apple's alternate feed URL forms, so you don't lose reviews. It retries rate limits and server errors a few times, honouring Retry-After. It never waits past the run's timeout: it stops early and keeps what it has.

What happens with a wrong app id or URL?

You get a free notice row naming the input and the reason, and everything else still runs. Nothing is charged for the skipped item. If nothing in the input is usable, the run returns only notice rows and charges nothing. It does not fall back to the example app. A malformed item from Apple is skipped the same way.

Use with AI agents / MCP

Add the actor to the Apify MCP server (mcp.apify.com). Claude, ChatGPT, Cursor and other MCP clients can then call it as a tool. Agents can ask questions like these:

• "What are people complaining about in the latest Duolingo release in the UK?" The agent sets apps, countries and ratings: ["negative"].
• "Where does Notion rank in US Productivity today?" The agent sets apps: ["Notion"], chartGenres: ["Productivity"] and chartTrackedAppsOnly: true. The run fetches charts only and costs about $0.0055.
• "Any new 1-star reviews for our app since yesterday?" The agent sets onlyNewReviews: true, ratings: ["1"] and its own stateKey.

Every field description in the input schema states what it costs. The output is small, flat and typed. Skipped items come back as free notice rows, so an agent can explain them instead of guessing. For exact targeting, agents should pass App Store ids or URLs rather than names.

Disclaimer

Unofficial. This actor is not affiliated with, endorsed by or sponsored by Apple Inc. App Store and iTunes are trademarks of Apple Inc. It reads only public, unauthenticated Apple endpoints. Review text belongs to its authors. Use the data in line with Apple's terms and applicable law.

Changelog

• 0.1 (2026-09-25): First release.
  • Metadata, including screenshots, advisories, the full description and English genre names in every storefront.
  • Reviews: dedupe, a new-since-last-run monitor that resumes after early stops (including a capped first run), star filter and per-version aggregates.
  • App names and search terms are searched in each storefront, with a check for Apple's stand-in results. Bundle ids are found in any of the listed storefronts.
  • Charts: 7 chart types that follow the App Store's Apps charts, genres and rank tracking. Up to 50 storefronts.
  • Recovery for Apple's cached empty review pages.
  • Unusable input returns free notice rows and is never replaced by the example app. A run without charged records costs nothing.