FitNS

FitNS is a Kotlin Android fitness app for nutrition tracking, strength training, body-weight progress, coaching signals, and optional synchronization with a private n8n instance.

All visible app UI copy is kept in US English.

• Dashboard with daily nutrition progress, hydration actions, training summary, and coach guidance.
• Nutrition logging with macro and micronutrient fields, meal filters, sectioned search, recent foods, favorites, custom foods, saved meals, barcode lookup wiring, label scan, and meal photo analysis entry points.
• Workout plans/templates backed by Room, including create, edit, delete, start, ordered exercises, default sets, rep ranges, and rest times.
• Active workout mode with per-set weight, reps, RPE, RIR, set type, completion status, previous performance, automatic rest timer, and finish/discard actions.
• Automatic personal-record detection for highest weight, reps at weight, estimated 1RM, and session volume.
• Body-weight logging with raw values, moving average, trend rate, and goal-distance summaries.
• Progress analytics for body weight, strength, muscle-group volume, and nutrition adherence.
• Optional Google sign-in state and local settings.
• WorkManager sync queue for local-first changes.

The app keeps the existing layered structure:

• core: auth, design system, settings, serialization, sync, and network primitives.
• data: Room entities, DAOs, migrations, mappers, local repositories, and remote n8n repository.
• domain: models, repository interfaces, calculators, and use cases.
• feature: Compose screens and ViewModels by product area.
• navigation: Compose Navigation graph and top-level app shell.
• di: Hilt bindings and app-level providers.

ViewModels primarily own UI state and events. Workout calculations such as double progression, estimated 1RM, active-session construction, PR detection, progress analytics, nutrition scaling, and daily coach readiness live in domain use cases.

• Kotlin
• Jetpack Compose
• Material 3
• Hilt
• Room
• DataStore
• Retrofit
• Moshi
• WorkManager
• CameraX
• ML Kit Barcode Scanning
• ML Kit Text Recognition
• Google Authentication
• Compose Navigation

Open the project in Android Studio and let Gradle sync, or run:

./gradlew assembleDebug

Release signing can be provided through environment variables or .env.android-signing. Do not commit secrets or private tokens.

Run the unit and Robolectric migration tests:

./gradlew testDebugUnitTest

Useful broader checks:

./gradlew lint
./gradlew assembleDebug

Current focused coverage includes meal scaling, daily coach readiness, nutrition adherence, body-weight moving averages, double progression, estimated workout progression suggestions, personal-record detection, and Room migrations through schema version 5.

The default configurable n8n base URL is:

https://pi.pufferfish-lenok.ts.net/

The app includes a Retrofit N8nApiService, settings for the base URL and bearer token, a connection test, barcode product lookup, meal-image analysis wiring, and a WorkManager-backed sync queue. Sync remains local-first and should tolerate offline usage.

CameraX and ML Kit dependencies are present. The barcode flow is wired from manual food entry to scanner to product lookup, with product-prefill support in nutrition logging. Follow-up hardening should expand empty, denied-permission, retry, unknown-barcode, and offline UX states.

Workout plans are persisted in Room as reusable templates. Starting a plan creates an active session in memory, prefilled from template defaults and previous performance. Finishing the session writes the workout, exercises, and sets atomically to Room and enqueues a sync payload.

Supported set types:

• Warm-up
• Normal
• Drop Set
• Failure
• AMRAP

The rest timer starts automatically when a set is completed and supports add/subtract time, pause, resume, and skip.

Nutrition supports manual food logging, meal type selection, favorites, recent-food ranking, custom foods with micronutrients, saved meals/recipes, meal scaling, copy-yesterday and copy-previous-meal actions, daily target tracking, hydration, label scanning, barcode lookup, and meal-photo analysis entry points.

Health Connect is not yet implemented. It should remain optional when added, with clean permission handling and no degradation for users who do not connect it.

Completed:

• Local-first Room repositories and schema migrations through version 5.
• Workout plan create/edit/delete/start.
• Active workout logging with editable sets, set types, RPE/RIR, previous performance, and rest timer.
• Personal-record detection from workout history.
• Domain use cases for workout progression, active-session construction, estimated 1RM, and PR detection.
• Nutrition search sections for recent, favorites, custom foods, and remote results.
• Saved meals, scalable recipes, and copy-previous meal actions.
• Progress analytics for body weight, strength, muscle-group volume, and nutrition adherence.
• Dashboard coach refactor into smaller explainable use cases.

In Progress:

• Barcode lookup UX hardening.
• Barcode scanner empty, retry, unknown-product, and offline state polish.

Planned:

• Health Connect import/export.
• Compose UI tests for workout and nutrition flows.
• Release R8/resource shrinking review and ProGuard hardening.