How aiappcritic works

More apps launch every week than anyone could ever try — more and more of them built with AI, and more and more of them sending what you type to an AI model. We do not think either is good or bad by itself. We think you deserve to know.

Curated, not crawled

Every app on aiappcritic was picked by a human. Developers can submit their apps, but nothing goes live without a review. We would rather show you 300 apps worth your time than 30,000 you have to dig through.

The AI fingerprint — two questions

AI is welcome here. We do not rank apps by how little AI they use — we answer two questions for every app: how was it built? and what does it do with AI — and with your data?

1 Built with AI

Core build — The code it runs on, the interface you touch, the words you read.

Code
Design & UI
Texts & docs

Around it — Store page, screenshots, ads and launch posts.

Marketing

By hand: Made without generative AI.
AI-assisted: AI helped (autocomplete, drafts, reviews), people wrote and decided the result.
AI-generated: Substantially generated by AI — and directed, reviewed and shipped by people.
Not disclosed: We do not know (yet).

The core build decides the label you see on every app. Marketing is shown openly, but does not change the label.

Hand-built: Code, design and texts made by people, without generative AI.
Hand-built with AI tools: Made by people, with AI assisting along the way.
Human × AI: One core part was generated with AI.
AI-led: Two core parts were generated with AI.
AI-built: Code, design and texts were generated with AI — directed and shipped by people.

Hand-built labels also say how well they are backed up: self-declared by the developer, evidence provided (public commit history, design files, build-in-public posts), or verified by us.

2 AI inside

Does the app itself use AI — on your device, in the cloud, or not at all? And if data leaves your device for an AI provider, developers must say which data: “the text you select”, “photos you tap Describe on”, “everything you type”. We also show which providers they use and whether your data trains AI models.

No AI inside — nothing you do goes to an AI model.
On-device AI / AI, no data shared — AI features, but your data does not go to AI providers.
Data to AI — some data goes to AI providers. The app page says exactly which.

Looking for apps that keep your data to themselves? Filter by No AI inside or No data to AI providers.

Human touch

A small meter shows how much human handwork went into the core build. Code counts 40, design 30, texts 20. Made by hand counts fully, AI-assisted 70 %, AI-generated still 30 % — because someone directed, reviewed and shipped it. Every app gets to tell why it used AI — or why not — in the developer's own words.

Hidden gems

Apps we or our users love (4/5 or more) that are still barely known: fewer than 500 App Store ratings and fewer than 500 GitHub stars. These are the ones that drown in the flood of new releases — we want you to find them.

Behind the scenes

Whenever we can, we add an interview, the origin story or an anecdote from development — for hand-built and AI-built apps alike. How someone vibe-coded an app in a weekend is just as interesting as a ten-year labour of love.

Early Access

Our beta corner for apps that are not finished yet. Developers share a public TestFlight link, a Google Play open test, a GitHub release page or a web app; you try it and answer three short questions — what you liked, where you got stuck, and one thing you'd change. Your answers go privately to the developer, reactions are public. There are no scores: unfinished apps deserve feedback, not verdicts. We check every link and every new version by hand, never allow direct downloads, and take builds down quickly when something looks wrong. When an app is released, it can graduate to a full entry. Try a beta.

Scores

App Store says — the average user rating on Apple's App Store, fetched through Apple's public lookup API, with the number of ratings so you can judge how meaningful it is.
GitHub says — stars of an open-source project, plus when the code last changed.
aiappcritic says — our own editor score out of 5.
Users say — the average of ratings on this site. We only show it once at least three people have rated.

Google Play and the Microsoft Store have no public API, so we link to them but do not show their ratings.

Comments & takes

Comments are limited to 200 characters. You can also share an X post or an article — we show a short summary and a link to the source. The first three posts of every new member are checked by a human. After that, you are trusted.

Sister site

Into games? aigamecritic does the same for indie games: curated, with an honest AI fingerprint for every single one.

How we make money

Some store links on this site may be affiliate links: if you buy through them, we may earn a small commission. It never costs you more and it never influences our picks or scores. We also show ads.