Hey everyone :wave: I wanted to share LAM, a film...
# feed
v
Hey everyone 👋 I wanted to share LAM, a film-inspired camera app for iOS and Android built with Kotlin Multiplatform and Compose Multiplatform. A common question around Compose Multiplatform is how far to push it before hardware constraints force you back to separate platform UIs (SwiftUI & Jetpack Compose). In LAM, I chose to share 100% of the interactive camera UI—mechanical dials, spring physics, and the camera state machine—in Compose Multiplatform, while keeping the pixel capture and GPU rendering on native pipelines. I’ve attached a simplified architecture diagram showing where I drew the boundary between shared Compose code and the native camera engines. What’s shared (commonMain): • Camera State Machine: roll management, shutter locks, exposure math, and C2PA provenance via Coroutines (
StateFlow
/
SharedFlow
). • Tactile UI: physical dials, animations, and haptic detents built once in Compose Multiplatform. What stays native: • iOS:
AVCaptureSession
+ Metal (
CVMetalTextureCache
mapping pixel buffers directly to GPU textures, zero Kotlin heap allocation). • Android: Camera2/CameraX + OpenGL ES (direct native memory access via JNI critical arrays). Stack: Kotlin 2.4.0 · Compose Multiplatform 1.11.1 · AGP 9.4.1 · Gradle 9.8.1 If you'd like to check out the UI interaction on real hardware: • iOS: apps.apple.com/app/id6773256954 • Android: play.google.com/store/apps/details?… For anyone building hardware-integrated apps with KMP: where do you draw the line between your shared Compose UI and platform-specific capabilities? Happy to discuss in the thread!
✅ 1
🎉 2