Let's Try

Around 10,700 lines of Swift across three targets: the phone app, a watch app, and widgets. SwiftUI and SwiftData, entirely local first, so there is no account and no backend. The interesting part is the scoring: instead of capping each pillar at 100, effort points are uncapped, so a punishing session actually outscores a short one. Habits are the core object, and the workout and cardio habits tick themselves from real logged activity.

Detail

• SwiftUI and SwiftData, local first. The only optional network call is AI food analysis, using your own key from the Keychain.
• Uncapped effort scoring: reps times effort times load, so heavy work is worth more.
• Ten data models covering habits, workouts, plans, day logs, food, and weekly check ins.
• Seeding repairs before it creates, after a shipped bug meant a plan that lost its exercises could never recover.