SparkyFitness — Self-Hosted AI-Powered Family Fitness Tracker

Introduction

SparkyFitness is a self-hosted fitness and health tracking platform built for families. It combines food logging, exercise tracking, water intake monitoring, and AI-powered coaching into a single app where household members share a dashboard and support each other's wellness goals.

What SparkyFitness Does

• Logs meals with nutritional data including calories, macros, and micronutrients
• Tracks workouts, steps, and exercise routines with progress charts
• Monitors daily water intake with reminders and goal tracking
• Provides AI-powered health coaching and personalized recommendations
• Supports multiple family member profiles on a shared self-hosted instance

Architecture Overview

SparkyFitness is a TypeScript application with a React frontend and a backend API. It integrates with LLM providers for AI coaching features. Health data is stored in a local database, ensuring all personal information stays on your own infrastructure. The Docker Compose setup bundles all components for simple deployment.

Self-Hosting & Configuration

• Deploy with Docker Compose using the included configuration
• Configure AI coaching by providing an API key for an LLM provider (OpenAI, Ollama)
• Create family member profiles through the web interface
• Set individual goals for calories, exercise, and hydration per user
• Data persists in the database volume with standard backup procedures

Key Features

• Family-oriented design with per-member profiles and shared dashboards
• AI coaching that analyzes habits and suggests personalized improvements
• Comprehensive food database for quick meal logging
• Visual progress charts and streak tracking for motivation
• Privacy-first: all health data stays on your own server

Comparison with Similar Tools

• MyFitnessPal — commercial app with ads and subscriptions; SparkyFitness is free and self-hosted
• Fitbit — hardware-dependent ecosystem; SparkyFitness works with manual input on any device
• Cronometer — nutrition-focused SaaS; SparkyFitness adds AI coaching and family features
• Wger — open-source workout manager; SparkyFitness includes nutrition, hydration, and AI coaching

FAQ

Q: Do I need a GPU for the AI coaching features? A: No. AI coaching uses cloud LLM APIs by default. You can optionally point it at a local Ollama instance.

Q: Can each family member have private data? A: Each member has their own profile. The shared dashboard shows aggregate family progress while individual data remains per-profile.

Q: Does it integrate with fitness wearables? A: Currently it relies on manual input. Wearable integrations may be added by the community.

Q: What nutritional database does it use? A: SparkyFitness includes a built-in food database and supports manual entry for custom meals.

Sources

• https://github.com/CodeWithCJ/SparkyFitness
• https://github.com/CodeWithCJ/SparkyFitness#readme