Short answer
A PWA is cheapest but has no real iOS store presence; a native wrapper gets your existing AI-built web app into both stores quickly with native features; React Native or Flutter give the best native experience but mean a second codebase. Coding agents make rebuilds cheaper, but they do not remove the cost of maintaining two products.
On this page
The question behind the question
When an AI-built web app gains traction, someone asks "can we have an app?". The answer depends less on technology preferences and more on three practical questions:
- Do your users need to find you in the App Store and Google Play?
- Which device capabilities does the core workflow genuinely require?
- Who will maintain a second codebase for the next three years?
With those answered, the choice between a PWA, a native wrapper and a cross-platform rebuild is usually clear.
The three options at a glance
| PWA | Native wrapper | React Native or Flutter | |
|---|---|---|---|
| What it is | Your web app, installable from the browser | Native app shell that loads your web app and adds native features | Separate mobile app with native UI components |
| App Store (iOS) | No direct route | Yes | Yes |
| Google Play | Yes, via Trusted Web Activity | Yes | Yes |
| Push on iOS | Only for installed web apps on recent iOS versions, with friction | Native push | Native push |
| Device APIs | Limited, varies by browser | Broad through native bridges | Full |
| Codebases | One | One web app plus a thin shell | Web app plus mobile app |
| Time to stores | Days (Android only) | Days | Weeks to months |
| Ongoing effort | Lowest | Low | Highest |
Option 1: Progressive Web App
A PWA adds a manifest and a service worker to your web app so users can install it to the home screen and use some features offline. AI builders such as Lovable or Bolt can add the basics in a prompt.
Strengths: no store review, instant updates, one codebase, no fees. Weaknesses: iOS users must discover the "Add to Home Screen" option themselves, there is no App Store listing, and capabilities on iOS lag behind Android. For internal tools where you can instruct users, or for Android-heavy audiences, a PWA can be enough.
Option 2: Native wrapper
A native wrapper is a real iOS and Android app whose main view is a WebView showing your web app, combined with native features such as push notifications, biometrics, file access, deep links and in-app purchases. Updates to the web app reach the mobile apps immediately.
For AI-built apps this route has a specific advantage: the agent or builder you already use keeps working on one codebase, and every feature you ship on the web appears in the apps. You do not have to teach the agent a second stack or keep two implementations in sync.
WebViewGold is the template we recommend here. It provides iOS and Android projects where you insert your URL, configure features like push, AdMob, offline mode and over 40 native APIs, and build store-ready apps without coding. The one-time purchase includes an approval guarantee (see their site for the terms), which matters because the main risk of wrappers is Apple's minimum functionality review. Our guide to Guideline 4.2 for web-based apps explains how to add enough native value.
Weaknesses: the experience is only as good as your mobile web app, and very interaction-heavy or offline-first apps will feel less native.
Option 3: React Native or Flutter rebuild
A cross-platform framework produces a separate app with native UI. Coding agents such as Claude Code, Codex or Cursor are now good at generating React Native and Flutter code, and can port screens from an existing React web app quickly. That lowers the initial cost considerably compared with a few years ago.
What it does not lower is the long-term cost. You now have:
- Two codebases whose features drift apart unless someone keeps them aligned.
- Two release processes, one of which goes through store review for every change.
- Native dependencies that need updates with each iOS and Android release.
- A shared backend whose API must stay compatible with old app versions still installed on phones.
Choose this route when the core experience depends on things a web view does poorly: offline-first data entry in the field, continuous background location, Bluetooth peripherals, camera processing, complex gestures or animations. Fully native Swift and Kotlin remain the choice for the most demanding apps.
apprat.io builds and maintains this kind of app with AI-assisted development plus human developer hours, which is a practical way to get a native app without staffing a mobile team. For agent-based workflows, see agentic coding tools for mobile app development.
A decision guide
Ask these in order and stop at the first "yes":
- Is the app internal, and can you instruct users how to install it? A PWA may be enough.
- Is the web app already good on phones, and are your must-have features notifications, sign-in, payments, sharing and camera uploads? Use a native wrapper.
- Do you need offline-first operation, background processing or deep hardware access? Rebuild with React Native or Flutter, or fully native.
- Unsure? Start with a wrapper. It gets you into the stores, gives you real usage data, and the backend work carries over if you later rebuild.
Common mistakes
- Rebuilding natively because "wrappers get rejected", when the real issue was a web app that was not mobile-ready.
- Choosing a PWA for a consumer app whose growth depends on App Store discovery.
- Letting an agent generate a React Native app without planning who maintains it, then falling behind on SDK and target API requirements; see maintaining an AI-built app after launch.
- Ignoring in-app purchase rules, which apply to wrappers and native apps alike.
Our recommendation for most AI-built apps
Start where your product already lives: on the web. Make it excellent on mobile browsers, ship it to both stores with WebViewGold, and measure. If usage shows that specific workflows need native depth, build those natively later with a clear business case. For a step-by-step conversion guide, read converting a web app into iOS and Android apps.
FAQ
Can a PWA be published in the Apple App Store?
Not as a PWA. To appear in the App Store you need a native app, which can be a wrapper around your web app or a native build.
Is a native wrapper slower than React Native?
Rendering inside a WebView depends on your web app's performance. A well-optimised web app feels fast; heavy animations and complex gestures generally feel smoother in React Native or Flutter.
Do AI coding agents make React Native the obvious choice?
They make the first version cheaper, not the maintenance. Two codebases still need parallel feature work, testing and store releases.
Can I start with a wrapper and switch later?
Yes. Your backend, accounts and store listings carry over. Many teams later replace the wrapper with a native build under the same app listing.