WorkoutTracker: AI-Powered Fitness Ecosystem 🏋️‍♂️🧠

WorkoutTracker is a state-of-the-art, fully native iOS & watchOS fitness application. It transcends traditional rep-counting by integrating Generative AI, Custom CoreML Computer Vision, and Biometric HealthKit Analysis to act as a truly intelligent personal trainer.

🔗 Part of a unified ecosystem: WorkoutTracker works in seamless synergy with its sister application, FoodTracker, providing a complete 360-degree view of your health, nutrition, and physical performance.

---

📸 Application Gallery

Dashboard & Biometrics	Neural AI Coach	Smart Anatomy Heatmap
Workout Hub	AI Program Architect	Legendary Routines
Advanced Stats & Radar	Progress & Predictions	Live Workout Analytics
Mechanics Breakdown	Extended Comparisons	Social Share Cards

---

🌟 Core Features

👁️ Custom CoreML & Computer Vision Tracking

• Action Classification Model: Uses a custom-trained .mlmodel integrated with an MLMultiArray prediction window (60 frames buffered). The AI natively understands if you are performing a rep or just resting.
• Live Pose Estimation: Utilizes VNHumanBodyPoseObservation to track skeletal mechanics and biomechanical angles (e.g., knee-to-hip depth) in real-time.
• Velocity Based Training (VBT): Detects if your bar speed drops significantly during a set and triggers auditory warnings before failure.
• Gesture Controls: Complete a set by showing a "Victory" (✌️) sign, or cancel tracking with an "Open Palm" (✋) using VNHumanHandPoseObservation.

🧠 Generative AI Neural Coach (Gemini API)

• Smart Program Architect: The AI analyzes your available equipment, days per week, and target muscles to instantly generate structured multi-day splits.
• In-Workout Proactive Adjustments: Chat with the coach mid-workout. Machine taken? Too heavy? The AI instantly swaps exercises or drops the load, natively altering the app's database state.

🍎 FoodTracker Synergy & Ecosystem

• Two-App Ecosystem: Deeply integrated with FoodTracker. Calories burned during heavy lifting in WorkoutTracker are instantly reflected in FoodTracker's daily allowance via Apple Health.
• Hydration Impact: WorkoutTracker reads water intake logged in FoodTracker to calculate your true CNS recovery score.
• Seamless Deep Linking: Jump instantly between tracking your macros in FoodTracker and hitting the iron in WorkoutTracker using custom URL schemes (foodtracker://).

⚡ CNS Fatigue & Biometric Analysis

• HealthKit Integration: Reads Sleep Analysis, Resting Heart Rate (RHR), and Heart Rate Variability (HRV) from Apple Health to calculate a real-time Central Nervous System (CNS) Index.
• Anatomical Heatmap: A custom SVG parser renders a highly detailed male/female muscular system. Muscles dynamically glow red/orange/green based on 48h-96h volume decay and RPE exhaustion.

📊 Advanced Data Analytics

• 1RM Prediction Engine: Uses Epley, Brzycki, and Lander formulas combined with linear regression to forecast your 1RM gains over 30/90 days.
• Symmetry & Radar Charts: Detects muscular imbalances (e.g., "Too much push, not enough pull") and renders beautiful Swift Charts to visualize mechanic breakdowns.

⌚ watchOS Companion App

• Real-time Sync: Uses WatchConnectivity to stream live Heart Rate to the iPhone dashboard and sync active workout states instantly.
• Standalone Tracking: Log sets, adjust weights via the Digital Crown, and finish workouts directly from your wrist while the iPhone stays in your locker.

🎮 RPG-Style Gamification & Achievements

• Leveling System: Earn XP based on workout volume and effort to level up your fitness rank (e.g., Novice, Iron Apprentice, Gym Titan).
• Dynamic Achievements: Unlock Bronze to Diamond tier badges for hitting milestones (e.g., 1000 tons lifted, 365-day streaks, early bird workouts).
• Streak Protection: Interactive mascot and push notifications warn you before neurological detraining begins.

🏝️ Live Activities, Dynamic Island & Widgets

• ActivityKit Integration: Live tracking of workout duration and active exercise directly on the Lock Screen and Dynamic Island.
• Interactive Widgets: Home Screen and Lock Screen widgets for quick access to your current streak, daily activity rings, and instant AI Coach status.

🌍 Production-Ready Localization

• Fully localized in English and Russian using Apple's modern .xcstrings catalog, dynamically adapting not just UI text, but AI Coach prompts and SVG anatomy labels.

---

🛠 Technical Architecture

This project strictly adheres to modern Apple development paradigms, prioritizing performance, safety, and scalable code design.

• UI/UX: 100% SwiftUI. Built with reusable components, custom ViewModifiers for Glassmorphism, 3D Parallax button effects, and rich Swift Charts.
• Architecture: Clean MVVM architecture combined with Apple's pure Observation framework (@Observable) for reactive state management.
• Data Persistence: SwiftData utilized with background @ModelActor operations. Heavy analytical calculations, JSON parsing, and DB migrations never block the Main Thread.
• Concurrency: Fully migrated to Swift 6 Strict Concurrency (async/await, TaskGroups, Actors, Sendable compliance), completely eliminating legacy GCD closures.
• Machine Learning Pipeline: Pure Vision framework for skeletal tracking + CoreML sequential windowing buffers to process time-series movement data smoothly.
• Media & Export: AVFoundation for camera buffer extraction, audio ducking, and Voice TTS. Custom VideoExportService to render dynamic heatmaps into MP4 shareable videos using AVAssetWriter.

---

🚀 Installation & Setup

To compile and run this project, you will need macOS Sonoma/Sequoia and Xcode 15.0+ (iOS 17.0+ Simulator or Physical Device required for SwiftData and Vision features).

1. Clone the repository:

   git clone https://github.com/Borisserz/WorkoutTracker.git

2. Configure the AI API Key: Locate Secrets.swift in the AppCore directory and insert your Google Gemini API Key:

   enum Secrets {
       static let geminiApiKey = "YOUR_GEMINI_API_KEY_HERE"
   }

3. Configure App Groups (Important for Widgets & WatchOS): Open WorkoutTracker.xcodeproj. In the Signing & Capabilities tab for all targets (App, Widget, WatchApp), ensure the App Group (group.com.yourname.WorkoutTracker) matches your Apple Developer account.

4. Build and Run: Select a physical iPhone (Highly recommended to test CoreML Computer Vision and HealthKit integration) and press Cmd + R.

---

📜 License & Copyright

Copyright (c) 2026 [Boris Serzhanovich]. All rights reserved.

This project is showcased for portfolio and demonstration purposes only. The source code, UI designs, and custom SVG anatomical models are proprietary. They are not licensed for public, commercial use, redistribution, or modification without explicit written permission from the author.