Flutter: if you want a pixel perfect version of your app that looks similar on Android and iOS. Some quite popular apps use it and make a shit ton of money, for example Headspace. More here: https://flutterhunt.com/

React Native: totally viable, because core business logic is the same. If you want native UI elements and all the iOS Liquid Glass stuff, but don’t want to have to completely separate code based, React Native is a viable option if you don’t mind installing lots of third party packages for functionality you get out of the box in Flutter or true native apps.

Native mobile: I have one smaller app with quite a custom calendar view, written with Claude Code natively for Android and iOS. Coding agents help and you get maximum control. Mentally, it’s a bit cumbersome to direct you agent to do everything twice. While one platform can serve as a reference implementation for the other, it’s kinda annoying mentally "to do the same feature all over again and test and direct the agent". But it’s doable.

KMP: No experience, I just hate most things around Android.

PWA: That ship has sailed. Nobody knows about PWAs, most PWAs are bad. It’s a bummer really.

Capacitor: make an SPA (offline-first preferably), and use native code plugins via Capacitor. Still quite feasible, comparable to Flutter in the sense that you need to build a custom UI or use a fake version of Liquid Glass and Material. I use Ionic for that, but it’s a bit dated (still on Material v2). Heavily depends on your skills as a web dev to get good UX.

I've built apps with millions of users using flutter. Quicker, faster and easier for building cross-platform apps if you know what you are doing.

This is coming from a native framework evangelist for many years, and then just tried flutter early days (which was far more difficult than how easy it is today). Now I write almost all my projects in flutter/dart where I can, especially with AI the speed from idea to production is insane.

We use flutter at work. When you need to target various platforms for the same code base it's extremely competent.

I’ve been a React Native developer for 8 years, I’ve written many apps with it professionally and personally. Until recently, I would have recommended it wholeheartedly.

This weekend I had an idea I set out to experiment with: if I have an existing React Native app, how much effort would it take an LLM to rewrite it into genuinely native apps. My idea being I no longer need to write cross-platform code and pay the intermediary framework costs, I’d get direct access to the underlying platform. Instead, I could just spec the features I need in the app and the LLM can build it.

It took me a year to build the original app to production quality by hand. My “twins”, as they’ve become known, were at 80% completion in three days of on-off work.

I’m pretty much convinced my experiment was a success. Which sorta terrifies me that I’ve specialised in a technology that can be, for all intents and purposes, be entirely erased.

The only caveat I can find is it is still a great technology for doing cross-platform mobile and web. Using something like Tamagui (as rough as that can be sometimes) allows genuine code sharing across all platforms.

Sucks most that I’m now actively looking for a new React Native role.

Every time we tried it, it resulted in a worse experience than we knew could be done using native stacks. It is like using sand as a foundation for your house. Random version updates break random things. As soon as you need something a bit advanced on the platform, you are stuck with some barely supported library. Easy to start, hard to maintain and hard to ship a solid product.