Kettlebell Workout Tracking App

Overview

A real-time kettlebell workout tracking application that uses computer vision to detect and analyze kettlebell exercises. Built with Preact, TypeScript, and TensorFlow.js for browser-based pose detection.

Key Features

• Real-time Pose Detection: YOLOv8 model for kettlebell exercise tracking
• Camera Integration: Live video feed with pose overlay visualization
• Workout Management: Start/stop workout sessions with proper resource management
• Responsive Design: Optimized for both mobile and desktop

Architecture

This project follows Domain-Driven Design (DDD) with Clean Architecture principles:

• Domain Layer: Core business entities and rules (Workout, WorkoutStatus)
• Application Layer: Use cases and services orchestrating business operations
• Infrastructure Layer: External integrations (camera, ML model, event system)
• Presentation Layer: UI components and hooks

Getting Started

# Install dependencies
pnpm install

# Start development server
pnpm dev

# Run tests
pnpm test

# Build for production
pnpm build

Project Structure

src/
├── domain/           # Business entities and types
├── application/      # Use cases and services
├── infrastructure/   # External adapters and event system
└── presentation/     # UI components and hooks

Documentation

• Architecture Decisions - Key architectural choices and rationale
• Development Guide - Coding conventions and workflow
• Domain Concepts - Business rules and domain model

Technology Stack

• Frontend: Preact with TypeScript
• Build: Vite with path aliases
• ML: TensorFlow.js with YOLOv8 pose model
• Testing: Vitest with Testing Library
• Package Manager: pnpm