PulseCoach

PulseCoach is an AI-powered adaptive fitness coaching app built with Flutter. It generates short, personalized daily micro-sessions (2–10 min) that adapt to the user's condition through on-device AI logic, sensor data, and contextual information (weather, exercise catalog).

Developed as the exam project for the DIMA (Design and Implementation of Mobile Applications) course at Politecnico di Milano.

📄 Design Document & Test Campaign — the full design rationale, architecture decisions, and test campaign report.

Concept

PulseCoach targets students and active people who want quick, decision-free workouts. The core differentiator: the AI isn't decorative — it runs on-device, adapts based on user feedback and sensor data, and explains its decisions.

Core Features

• Profile & Goals — fitness level, objective (cardio / strength / mobility / well-being), available time, and constraints
• AI-Generated Daily Plan — three micro-sessions per day, dynamically adapted after feedback (RPE 1–10), sensor data, and weather
• Session Catalog — a curated exercise library with an offline fallback
• Progress Tracking — session history and completion trends
• Social — an opt-in accountability layer
• Wear OS companion app — session and rest display synced from the phone
• Full light/tablet responsive layout with rotation support, and English/Italian localization

Tech Stack

• Flutter / Dart, flutter_bloc for state management, go_router for navigation
• Drift (SQLite) for local persistence, get_it + injectable for dependency injection
• Supabase for auth, storage, and backend functions
• On-device AI logic for daily plan generation, running in a background isolate
• Sensors & health data via sensors_plus, health, geolocator
• Wear OS module (pulse_coach/wear) with phone↔watch communication

Project Structure

Flutter_PulseCoach/
├── docs/                          # Design docs, requirements, presentation
│   └── PulseCoach Design Document & Test Campaign.pdf
├── pulse_coach/                   # Flutter app
│   ├── lib/
│   │   ├── ai/                    # On-device AI / behavioral state machine
│   │   ├── core/                  # Cross-cutting concerns (DI, DB, error handling)
│   │   ├── features/              # Feature modules (onboarding, today, sessions_catalog,
│   │   │                          #   progress, social, settings, subscription, ...)
│   │   └── shared/                # Shared UI components
│   ├── test/                      # Unit, widget, bloc, and integration tests
│   └── wear/                      # Wear OS companion app
└── supabase/                      # Backend functions, migrations, and tests

Getting Started

cd pulse_coach
flutter pub get
flutter run

Run the test suite:

cd pulse_coach
flutter test

Documentation

• docs/PulseCoach Design Document & Test Campaign.pdf — full design and test campaign document
• docs/REQUIREMENTS.md — DIMA exam requirements
• docs/FLUTTER.md — proposal & compliance analysis
• docs/demo-runbook.md — demo runbook