Inspiration

Quality personal training is expensive, often costing over $100 an hour, leaving tailored fitness guidance out of reach for many. Furthermore, most static workout applications fail to accommodate physical limitations, pre-existing injuries, or equipment constraints. FitForge was inspired by the goal to democratize adaptive health coaching—creating an intelligent, injury-aware personal trainer that brings personalized fitness and gamified bodyweight progressions directly to any browser for free.

How We Built It

FitForge was engineered over a 4–5 month development and testing period with a robust full-stack architecture:

Backend Framework: Built with C# and ASP.NET Core, managing core business logic, custom REST APIs, and authentication flows.
Database Infrastructure: Utilizes a structured MySQL relational database hosting user profiles, active workout sessions, and multi-session data persistence.
AI Intelligence: Integrated with the Gemini API to power an interactive coach that dynamically tailors training routines around specific fitness goals, daily availability, and strict injury protocols.
Frontend & Deployment: Features a dark-themed, responsive web UI with live session logging, a gamified calisthenics skill tree, and cloud deployment hosted live on Render.

Challenges We Faced

Mathematical Progression Modeling: Designing an adaptive algorithm to dynamically compute progressive overload metrics and volume adjustments required balancing variable load equations (\text{Total Volume} = \sum_{i=1}^{n} (\text{Sets}_i \times \text{Reps}_i \times \text{Weight}_i)) across training sessions:

$$ \text{Total Volume} = \sum_{i=1}^{n} (\text{Sets}_i \times \text{Reps}_i \times \text{Weight}_i) $$

Contextual AI Prompting: Structuring prompt contracts so the AI coach consistently adheres to strict injury-safe modifications without generating redundant exercise recommendations.
State & Session Persistence: Managing real-time active timers and drop-set counters across web sessions while ensuring seamless synchronization with the MySQL backend.

What We Learned

Full-Stack C# Integration: Gained deep experience in structuring clean ASP.NET Core controllers, handling asynchronous API requests, and managing relational database schemas.
Prompt Engineering for Health Apps: Learned how to enforce strict safety constraints within LLM prompt chains to ensure generated workout routines prioritize joint safety and progressive fatigue management.
User-Centric Web Design: Discovered how gamification elements—such as calisthenics skill trees, milestone badges, and streak counters—significantly boost user consistency and long-term engagement.