An AI-powered fitness assistant that creates customized workouts, macro-based nutrition plans, and recovery-focused sleep routines tailored to your daily readiness and fitness goals.
The Wellness Companion is a standalone, high-performance Agent Service Provider (ASP) acting as an intelligent, coordinated personal training and recovery assistant.
The backend deploys four specialized sub-agents coordinating under a central orchestration framework:
Sleep Coach (Circadian Optimizer):
Analyzes sleep duration, quality, caffeine logs, and blue-light exposure.
Generates custom, hour-by-hour bedtime routines and behavior schedules.
Fitness Coach (Dynamic Trainer):
Formulates exercise plans based on experience level (Beginner, Intermediate, Advanced).
Automatically adapts routines based on joint pains, muscle tightness, or high fatigue signals.
Nutrition Coach (Performance Dietitian):
Computes daily target calories and macros (Protein, Carbs, Fats) tailored to fitness goals.
Designs customized daily meal structures respecting exclusions (e.g. vegetarian, vegan, allergies).
Nexus Orchestration Engine (The Supervisor):
Evaluates combined daily sleep and stress logs.
Dynamically triggers safety alerts and forces other coaches to adapt (e.g. replacing a high-intensity workout with active recovery and boosting recovery-aiding macros if sleep hours are low or stress levels are elevated).
Developed a clean, standalone TypeScript/Express backend with a multi-agent consensus architecture (Sleep, Fitness, Nutrition, and Nexus agents) that dynamically adjusts routines based on daily recovery metrics.
Implemented a fully compliant OKX.AI Model Context Protocol (MCP) gateway allowing automated tool discovery and execution.
Migrated the entire codebase to the high-speed Groq SDK (groq/compound), set up Ngrok secure tunnels, and successfully deployed the backend side-by-side on an AWS EC2 instance using PM2.
None