Back to projects
Active

HealthPulse

Personal health dashboard integrating Apple Health with AI coaching and gamification

Next.jsReactTypeScriptSQLiteClaude API

A Mac-based health tracking dashboard that integrates Apple Health data with AI coaching, gamification, and activity analysis. Built entirely using Claude Code without writing any code manually.

The Problem

Apple Health collects extensive data but presents it in fragmented views. I wanted a unified dashboard that could show trends, correlations, and provide actionable insights—without sending my health data to the cloud.

The Solution

A local-first health dashboard that:

  • Imports Apple Health data via XML export
  • Visualizes trends across weight, activity, and cardiovascular metrics
  • Provides AI coaching with personalized recommendations
  • Gamifies progress with XP, levels, and achievements

Core Features

Health Tracking

  • Weight monitoring (manual or imported from Apple Health)
  • Activity integration (steps, workouts, flights climbed)
  • Body composition tracking (body fat %, lean mass)
  • GPS route visualization for walks and runs

Advanced Metrics

  • VO2Max cardiovascular fitness tracking
  • Resting heart rate trends
  • Heart rate variability for recovery insights

AI Coaching

An AI Coach powered by Claude analyzes complete health history to provide personalized recommendations. The coach assesses recovery status based on physiological data and suggests training modifications.

Gamification System

  • XP rewards (10 points for logging weight, 15 for hitting step goals, 20 for workouts)
  • Progressive leveling (Levels 1-50+)
  • Streak tracking with bonuses
  • Achievement badges for milestones
  • Daily quests for engagement

Data Visualization

Interactive multi-axis charts with customizable timeframes (30D/90D/1Y/3Y/5Y/10Y views), trend analysis showing correlations between activity and weight changes, and GitHub-style activity heatmaps.

Privacy First

All data is stored locally in SQLite—no cloud synchronization, analytics, or telemetry. The AI Coach only receives data when explicitly queried. Your health data stays on your machine.

Built with AI

This project represents approximately 8,700 lines of TypeScript across 21 commits, completed in roughly 20 hours over three weeks—entirely generated by Claude Code based on description and iteration feedback.

No manual coding required. Just conversation with an AI about what I wanted to build.