App Store Review Data API Access: What You Can (and Can't) Get in 2026

App store review data is the collection of ratings, written reviews, and related metadata (date, app version, store country, developer replies) that users leave on platforms like the Apple App Store, Google Play, and others. It's used by product teams, marketers, and analysts to track sentiment, surface bugs and feature requests, and benchmark against competitors.

Quick answer: what can you access?

Officially, only your own app's reviews, and even that comes with gaps (no rating-only feedback, limited history, no global endpoint).

Anyone else's reviews – competitors, a category, the whole market – have to come from capped public feeds or aggregated through a third-party provider, each with its own coverage characteristics.

A typical review record includes: star rating, review text, a reviewer nickname, date, app version, store country, and any developer reply. What it almost never includes: who the reviewer actually is. No verified identity, demographics, or person-level location.

Monitoring your own app's reviews: what the official APIs cover

If you're a product or growth team monitoring your own app, start here.

Each platform's official API works, but each has a big blind spot worth knowing before you build a workflow around it.

App reviews are one of many signal sources teams are putting to work, this roundup of social data use cases in 2026 covers the others.

Apple App Store Connect API

The App Store Connect API lets you pull and reply to your own app's reviews, but only for apps your account already controls. There are essentially three gaps to plan around:

• No average rating endpoint. There's no API call that returns your current star average, you'd need to calculate it yourself from raw reviews.
• Star-only ratings are invisible. Because users can rate without writing text, the reviews endpoint undercounts your real rating volume versus what shows in App Store Connect.
• No global view. There's no single worldwide call, you pull country by country, across 116 storefronts.

💡 If you're tracking sentiment trends, don't anchor to "average rating" from the API, it's not there. Build your sentiment metric from the review text itself.

Google Play Developer API

The Reply to Reviews API covers production versions of your own app only. Two things will trip up a new integration:

• Text-only reviews. A rating with no written comment doesn't appear via the API at all.
• One-week window. The API returns only reviews from the last seven days. For full history, you're exporting a CSV from the Play Console by hand.

💡 If this API is your only pipeline, you need a job that runs at least weekly, or you will silently lose reviews.

Microsoft Store API

The most generous first-party access of the three. The analytics API returns your own app's reviews across any date range, filterable by market and device. The catch, though, is that the companion respond-to-reviews API is currently documented as broken (as of June 2026), with developers redirected to Partner Center.

The takeaway

Apple, Google, and Microsoft each handle their review APIs differently, but they share one thing: official access stops at the edge of your own app.

None of these APIs has an endpoint for "show me reviews of app X" where X isn't yours. There's no parameter you're missing, no hidden permission to request, no enterprise tier that unlocks it. It's structurally absent from the API design itself.

Each platform's developer program is built around the relationship between you and your own listing: replying to your reviews, tracking your ratings, managing your own app's presence.

So if your use case is "how is our app doing," the official APIs (with the gaps above) are the right starting point. If your use case is "how is our category doing" – competitor sentiment, market-wide trends, benchmarking – the official APIs are not going to answer that question, no matter how the request is structured.

If "build it ourselves vs. find another way" is a familiar fork in the road, this breakdown of the build vs. buy decision for social data pipelines walks through how that calculation plays out.

That's a signal that you're asking a different kind of question, one that needs a different kind of data source. ⚠️

Can I get a competitor's App Store reviews?

Not through any official API, that data is only accessible through public feeds (with strict caps) or third-party providers that aggregate it. Apple, Google, and Microsoft all restrict their review APIs to apps you control, with no parameter, permission, or paid tier that unlocks competitor data.

Competitor and market app review data: what's available from public feeds

The instinct, once the official APIs run out, is to treat "public" as a workaround – if a review is visible on the App Store or Play Store page, surely there's a way to pull all of it.

There usually is a way to pull some of it. Each platform exposes a public endpoint of some kind, and each one comes with a specific, documented ceiling on how much you actually get back – a cap on volume, a window on recency, or in one case, almost nothing at all. ⬇️

Getting broad review coverage sounds like a given, but it isn't.

A single public feed runs out fast: Apple's RSS feed caps at roughly 500 recent reviews per app, full stop, no matter how often you check it.

So any data source can truthfully say "yes, we cover that app," because everyone is drawing from the same capped pool. The question that truly tells you something is how they get past that ceiling.

➡️ Do they pull from multiple feeds and methods? Do they capture reviews before they age out, so history doesn't just evaporate?

That's the difference between a dataset that's just reflecting the same 500 reviews back at you, and one that's genuinely broader than what any single public endpoint can offer.

This isn't unique to app reviews, structural data incompleteness shows up across most platforms, and the majority of teams don't find their gaps until an insight comes back wrong.

Is it legal to collect App Store review data?

Yes, collecting publicly available review data has solid legal precedent, but it comes with conditions.

There's real precedent that collecting publicly available data can be defensible.

But "publicly available" isn't the end of the conversation. Platform terms place restrictions on automated access, and reviews routinely contain personal data, which means GDPR and CCPA considerations apply regardless of how the data was collected.

In practice, this isn't a reason to avoid the data, it's a reason to be precise about governance: how it's collected, how personal data is handled, and how the resulting dataset is described to the people using it.

App store rating bias: why star ratings are misleading

App Store star ratings are misleading because they reflect who chooses to leave a rating, not how the overall user base actually feels – research shows the gap between the two can be large.

Even when you can see the reviews themselves, the star rating sitting next to them is doing less than it looks like it's doing.

It feels like a clean summary – thousands of people, averaged into one number, surely that's the closest thing to ground truth you'll get.

But that number is shaped by who bothers to leave a rating at all, not just how people feel about the product, and the gap between those two things is large enough that researchers have spent years studying it. ⬇️

J-shape distribution: why most reviews are 1 or 5 stars

Most app ratings cluster at the extremes – 1 star or 5 stars – because the people most likely to bother rating an app are those who loved it or hated it, not the majority who felt "fine" about it.

Online ratings follow what researchers call a J-shaped distribution.

Two self-selection biases cause it: acquisition bias (mostly people predisposed to like a product use it at all) and underreporting bias (people with extreme experiences are far more likely to post than the merely satisfied).

Together, they make the average rating a biased estimate of real quality.

The scale of the bias is striking: in a study where everyone was made to write a review regardless of how they felt, ratings came out roughly normal, only ~3% gave top marks and ~7% gave the lowest. The skewed public distribution reflects who bothers to post, not how people actually feel.

How in-app review prompts skew your app store rating

Ratings are also shaped by when developers ask.

Apple recommends prompting right after a user completes a task, and the entire design is built to catch happy users who'd never otherwise leave a review, pulling the average up.

App developers treat this as a tunable lever: smart prompt-timing can add half a star to a full star, and since apps under 4.0 are effectively invisible in competitive categories, the incentive to optimise this is strong.

💡 When benchmarking against competitors, the star average is comparing two things at once: genuine sentiment, and how aggressively each developer optimises their rating prompt. Read the distribution and the text. Don’t treat the headline number as signal.

What reviews are useful for (and where the hype overshoots)

Strip away the star rating, and what's left (the actual written text) is where the real value lives. ⬇️

• App reviews are a recognised source for requirements engineering, meaning teams have built entire methodologies around mining them for bug reports, feature requests, and even less obvious needs like performance complaints and security concerns
• At least 182 academic papers on app-review mining were published between 2010 and 2020 alone.

But there are limitations.

When researchers went back and tried to reproduce the results of popular opinion-mining techniques, the accuracy didn't hold up — the methods performed worse than the original papers claimed. The reason isn't surprising once you think about it: reviews are short, written in a rush, full of typos and slang, and people often bury the actual point in a sentence about something else entirely. That's a hard input for any automated system to parse cleanly.

Using AI to analyze app reviews: why data prep matters

If you’re thinking you’ll just feed your reviews into an AI and it'll spit out your product roadmap, the problem is everything that happens before and after the AI step.

Raw reviews are duplicated, scattered across versions and countries, and full of noise that makes it hard for any model to separate a real pattern from a handful of loud outliers.

Feed that in directly and you'll get a confident-sounding summary that's wrong in ways that are hard to spot.

➡️ Our recommendation? Clean and structure the data first, then have someone who knows the product sanity-check what comes out the other end.

The stakes get even higher in regulated contexts. Our post on adverse event monitoring from social media shows what happens when that same text-analysis pipeline has to meet pharmacovigilance standards.

The scale of App Store review fraud: how many reviews are fake?

Apple classifies roughly one in nine new reviews as fraudulent.

To put that in perspective, in 2024 Apple handled over 1.2 billion ratings and reviews and removed more than 143 million of them as fake; and that number climbed to around 195 million removed in 2025, a scale one report described as an industrial underground market for reviews.

And this isn't just about a few apps looking better than they deserve. It's competitive: fake reviews push apps higher in the rankings, which means real users end up downloading something that only looks popular, at the expense of apps that earned their position.

💡 If you're pulling review data from public feeds without any filtering or deduplication, think of it this way: roughly 1 in every 9 reviews in that dataset could be fake, and that's just Apple's own detection rate for what they catch (the true number, including what slips through, is likely higher).

Whatever you build on top of that data – sentiment tracking, feature-request analysis, competitor benchmarking – inherits that. A pattern that looks like "users really want X" might partly be "a bot campaign said X." Cleaning this out is a crucial step.

App store AI review summaries: why they're a black box

Both Apple and Google now show AI-generated summaries of app reviews. But neither discloses how they're built, which reviews they weight, or what gets left out, making them unreliable for serious analysis.

In 2025, both major stores inserted a new layer between reviews and readers.

Apple added LLM-generated review summaries in iOS 18.4, refreshed weekly for apps with "enough" reviews, though the threshold isn't disclosed. Google Play followed in late 2025 with "Users are saying" summaries and topic chips.

Convenient for shoppers, but useless for analysis. There’s a real lack of transparency that means you can't see how it was built, which reviews it weighted, what it dropped, or why your app doesn't have one yet.

App review data access: a quick can/can't checklist

‍✅ You can:

• Pull your own apps' reviews from each store's official API (with the gaps above)
• Get a recent, capped sample of any iOS app's reviews via Apple's public RSS feed
• Read Steam reviews fairly completely
• Buy aggregated, cross-platform market data from a third-party provider

❌ You can't (via official, first-party means):

• Get a competitor's full review history
• Pull complete history without a manual export
• See star-only ratings or any real reviewer identity
• Treat public feeds as a complete or fraud-free dataset on their own

Once you know what's accessible, the next question is whether it's enough: this framework for evaluating data coverage applies just as well to review data as it does to social.

What this means for your team

Everything we’ve covered above all turns out to be the same constraint wearing different outfits:

1. Official APIs only show you your own app
2. Public feeds cap out fast
3. Doing this properly takes real volume
4. And a chunk of what's out there is fraudulent

Different symptoms, same root cause – getting a complete, trustworthy picture of app review data takes more than a single API call.

If your need is narrow (just your own app, just recent reviews, just for monitoring) the official APIs are fine. Slower than you'd like, and missing pieces here and there, but workable for that job.

The moment your need grows – you want to see what competitors' users are saying, you need history that goes back further than a week or a few hundred reviews, you're covering multiple countries, or you want data clean enough to actually run analysis on or feed into a model – the picture changes.

None of this is impossible to do yourself, but it's also not a side project. It means pulling from more sources than any single feed offers, handling personal data the right way, filtering out the roughly 1-in-9 reviews that Apple itself flags as fake, and then doing it all again every time a platform tweaks its API or feed structure.

That combination is exactly what makes review data work better as infrastructure than as a one-off project. Collected broadly, deduplicated, filtered for fraud, and delivered through a single API in a consistent schema, so the time your team spends goes into the analysis itself, not into rebuilding the pipeline every few months.

Datashake delivers social and review data as clean, normalised infrastructure through a single API, so product, intelligence, and CX teams can build on the signal instead of collecting it. Find out more

Frequently Asked Questions

Can you scrape App Store reviews legally?

Collecting publicly available data has solid legal precedent, but platform terms place restrictions on automated access, and reviews often contain personal data subject to GDPR and CCPA. What matters most is governance, how the data is collected, handled, and described to the people using it.

How accurate are App Store ratings?

Less than they appear. Star ratings follow a J-shaped distribution driven by who chooses to rate at all (mostly people with strongly positive or negative experiences) not a representative sample of all users. In-app prompt timing also skews ratings upward by capturing happy users who wouldn't otherwise leave a review.

Can I see a competitor's App Store or Google Play reviews?

Not through any official API, these are restricted to apps you control. You can get a capped, recent sample (around 500 reviews) of any iOS app via Apple's public RSS feed, or work with a third-party provider for broader, aggregated coverage.

How many App Store reviews are fake?

Apple itself classifies roughly 1 in 9 new reviews as fraudulent, removing well over 100 million fake reviews annually. This means any review dataset pulled without filtering or deduplication likely contains a meaningful share of fraudulent entries.

Can AI accurately summarize app reviews?

The underlying signal in app reviews is useful for spotting bugs, feature requests, and trends. But raw reviews are messy, duplicated, and inconsistent, so reliable AI analysis depends on clean, structured input and human review of the output, not a single prompt.

‍