Caffy VS JavaScript Tutorial

Caffy's answer

Most caffeine trackers are calculators: you log a drink and they add up milligrams. Caffy models what your body is actually doing with it.

• It uses the Bateman equation to model absorption and elimination together, so the curve rises for roughly the first 45 minutes and then decays, instead of dropping the moment you log a drink.
• It calculates a personal half-life rather than assuming five hours for everyone. Pregnancy, smoking, oral contraceptive use, age and sex all shift that number, and between two people it can differ by a factor of six.
• Everything else follows from that: your daily cutoff time, your peak focus window, and a nightly sleep score estimating how much caffeine will still be circulating at bedtime.

The model and its sources (FDA, EFSA, NIH, Sleep Foundation) are documented inside the app.

Caffy's answer

• The number is yours, not an average. Competitors apply one half-life to everyone. Caffy asks about pregnancy, smoking, contraceptive use, age and sex, and recalculates from there.
• It answers the question people actually have. Not "how much did I drink" but "can I still have this coffee and sleep tonight".
• It lives in the OS. Home Screen and Lock Screen widgets, Live Activities, a Control Center control and Apple Health integration, so you see your level without opening the app.
• It works offline. Every entry is written locally first and synced afterwards.
• It is properly localized. English, Turkish, German, Spanish and Japanese, all hand written.

Caffy's answer

Three groups, in order of size:

• People whose sleep is worse than it should be. They suspect caffeine and want to see the connection instead of guessing. This is the core audience.
• People with a medical reason to watch intake. Pregnancy, arrhythmia, anxiety, reflux, or a doctor's instruction to stay under a specific milligram limit.
• People cutting down or quitting. They use the reduction programs to taper without withdrawal headaches.

Skewed toward iPhone users in English speaking markets, roughly 25 to 45, comfortable with health data but not necessarily technical.

Caffy's answer

I built it because I could not sleep and did not believe caffeine was the reason. My last coffee was in the early afternoon, which every article says is fine.

When I actually worked through the pharmacokinetics, the "stop at 2 PM" advice turned out to be an average dressed up as a rule. It assumes a five hour half-life and a midnight bedtime, and almost nobody matches both. My own cutoff was hours earlier than I had assumed.

No app would tell me my number, so I wrote one. It started as a single decay curve and grew into the sleep score, the cutoff time and the reduction programs as I kept running into the next question.

Caffy's answer

• iOS app: Swift and SwiftUI, MVVM with the Observation framework, SwiftData for the offline first local cache
• Charts: Swift Charts
• System integration: WidgetKit for Home Screen, Lock Screen and Control Center, Live Activities, HealthKit, UserNotifications, BackgroundTasks
• Backend: Supabase (PostgreSQL with row level security, anonymous auth, Edge Functions)
• Subscriptions: RevenueCat
• Analytics: PostHog
• Localization: hand written translations for English, Turkish, German, Spanish and Japanese