How We Think.
How We Build.

AI photo editing apps were trending on the App Store. Most were either too complex or too slow. We saw an opportunity for a one-tap, instant-result experience with no learning curve.

Day 1: Idea locked. Days 2–4: Claude + Cursor built the core editing pipeline in SwiftUI using CoreML for on-device processing. Day 5: UI polish and App Store assets. Day 6: Submitted.

We chose on-device AI over cloud for zero latency. RevenueCat paywall on Day 3. We launched with 3 filters instead of 20 — the minimal viable delight.

Calorie tracking apps all require manual logging. Users quit within 3 days. We wanted to make logging as fast as taking a photo — zero friction, maximum insight.

GPT-4o Vision identifies food from a photo and returns macros in under 2 seconds. We built a minimal SwiftUI interface around this single moment: snap, see, know. Everything else is secondary.

We cut the history screen on day 3 — shipping was more important. We added it in Week 2 after launch. Core loop first, features later. This is the Kodeora way.