FitSync – Personal Health & Fitness Tracker

Project Codename: FitSync
Current Stage: Early MVP / Prototype (Health Dashboard + Planned ML Posture Module)
Platform: Flutter (Android)
Last Updated: February 2026
Primary Data Source: Sahha Flutter SDK (Passive Health Sensing)
Planned ML Extension: On-device Pose Estimation & Posture Feedback
Backend: Firebase Authentication + Firestore
UI Style: Dark-themed, modern, green accent (#7CBA3B)

---

1. Project Goal

Build a modern, privacy-first daily health companion that:

• Passive Tracking: Collects steps, heart metrics, sleep, energy, and exercise via Sahha.
• Visual Insights: Provides clean daily summaries, progress visuals, and persistent goals.
• Gamification: Rewards consistency with XP, levels, and streak celebrations.
• Active Coaching (Future): Uses on-device machine learning to detect posture in real-time during sitting, yoga, or exercise, evolving the app from a tracker to a form coach.

---

2. Current Core Features (Implemented)

Feature Area	Status	Implementation Notes
Sahha SDK Integration	✅	Sandbox env; 20+ sensors (Steps, HR, HRV, Sleep, Energy).
Daily Health Dashboard	✅	Activity rings, goal progress bars, sleep & exercise lists.
Persistent Daily Goals	✅	Automatically carries over goals from the most recent entry.
Step-Goal Reward	✅	+100 XP & full-screen StreakAnimation on goal completion.
Date Browsing	✅	table_calendar integration for historical data viewing.
Heart Rate Detail	✅	Dedicated HeartRateGraph route.
Firebase Sync	✅	Auth and per-user Firestore collections for goals/gamification.
Metric Estimation	✅	Cal ≈ steps × 0.03; Distance ≈ steps × 0.0008 km.

---

3. Planned ML / Posture Detection Module

Objective: Real-time form correction for strength training, yoga, and desk-work posture.

Technology Strategy

We will utilize google_mlkit_pose_detection for the 2026 MVP due to its stability and lightweight performance on mobile.

• Landmarks: Detect 33 body points (shoulders, hips, knees, etc.).
• Logic: Calculate joint angles (e.g., knee-hip-shoulder) to classify form quality.
• Feedback: Real-time visual overlays + haptic/voice cues.

Logic Example: If $KneeAngle < 80^\circ$ or $BackAngle > 160^\circ$ during a squat, trigger a "Straighten Back" alert.

---

4. Firestore Data Layout

users/{userId}/
  ├─ healthGoals/{yyyy-MM-dd}
  │    ├─ stepGoal, activeMinutesGoal, activeCaloriesGoal
  │    └─ steps, activeCalories (daily snapshot)
  ├─ gamification/data
  │    └─ experience, level, totalPoints, streak, achievements
  └─ postureSessions/{sessionId} (PLANNED)
       ├─ date, exerciseType, durationSeconds
       ├─ avgPostureScore: 0–100
       └─ alerts: [{timestamp, message, severity}]