EppleyWorkoutConsulting

Inspiration

Personal trainers cost $50–100/session. Most people can't afford personalized coaching, so they either follow generic YouTube workouts or quit entirely. We wanted to make AI the great equalizer — giving everyone access to a plan that actually fits their body, goals, and schedule

What it does

Eppley Consulting is a mobile fitness app that generates fully personalized, multi-week workout and nutrition plans using AI.

Users complete an 8-step onboarding (weight, age, goals, experience level, diet preferences, days available) GPT generates a 4-week progressive plan with daily workout blocks, nutrition targets, and recovery days Plans are cached locally so you can access them offline anytime The app adapts to whether users want a diet plan or just workouts

How we built it

Frontend: Flutter (iOS-first, Cupertino design, Provider state management) Backend: FastAPI + Python, hosted with Uvicorn Database: PostgreSQL via Supabase AI: OpenAI API with structured outputs (responses.parse()) — guarantees the AI returns exactly the JSON schema we need, no hallucinated fields Auth: Device UUID-based authentication for frictionless mobile onboarding

Challenges we ran into

Getting OpenAI to reliably return structured multi-week plan data required careful Pydantic schema design and using the structured outputs API instead of plain prompting Syncing the Flutter data models with backend Pydantic schemas without type mismatches across the client-server boundary Designing the onboarding UX to collect all necessary data without overwhelming the user

Accomplishments that we're proud of

AI-generated plans that actually feel personalized — not just template swaps Clean full-stack architecture from Flutter → FastAPI → Supabase with consistent type safety end-to-end Offline-first approach with local plan caching means the app is useful even without connectivity

What we learned

OpenAI's structured output API is a game-changer for AI apps that need reliable data formats Device-level UUID auth significantly reduces signup friction on mobile Flutter's platform-aware widgets make it easy to ship a native-feeling iOS app fast

What's next for EppleyConsulting

Progress tracking & adaptive plan updates based on completed workouts Apple Health / Google Fit integration Social features: share your weekly plan, compare with friends Wearable integration (Apple Watch) for real-time workout guidance

Built With

• dart
• fastapi
• flutter
• jwt
• openai
• postgresql
• python
• supabase