Odinnou

1. Frontend: Kotlin Multiplatform + Compose Multiplatform, shared UI across iOS/Android.
   Kotlin was hard to love at first the module/folder hierarchy fights you early on. But once it clicks, KMP is a banger on almost every level: the IDE, the language itself, and the ability to drop down to native when you actually need to.

2. We went full Compose Multiplatform for the UI... except one screen: the audio player.
   Native audio APIs on iOS/Android are just too different (and too finicky) to abstract cleanly, so that one component is native on each platform. Everything else one codebase.

3. Backend: .NET 10, hosted on Azure Container Apps.
   Why .NET over Java or Go? Best perf/productivity tradeoff out there right now. C# is more expressive than Go, and the ecosystem is lighter than Java's (minimal APIs, EF Core, native DI = way less boilerplate). vs Go specifically: we keep strong typing and better abstractions as the codebase grows, without giving up performance .NET 10 is now neck and neck with Go in benchmarks.

4. Why Azure specifically: it's the only cloud that gives you a free Postgres DB with basically no limits to get started. For a bootstrapped build, that's a big deal real Postgres, zero cost, day one.

5. Dev setup: Rider on Apple Silicon. Excellent tooling, natively optimized for ARM, no Rosetta tax.

6. TL;DR: shared UI via KMP + Compose (native only where it truly matters), a .NET 10 backend that punches above its weight, and Azure Container Apps + free Postgres to keep costs at zero while we validate.

   Building in public questions/roasts welcome

What s more important during Caresse audio session?

• Personalization

• Voice quality

• Story

• Sound design

• ?

Which feature would you add next to Caresse?

• Long-distance mode

• More voices

• Apple Watch support

• ?