FlexFit AI

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments thatProject Name: FlexFit AI

An adaptive AI wellness and pantry assistant that recalibrates workout schedules guilt-free, offers AI coaching, and scans receipts to track nutrition.

Inspiration: Many fitness apps penalize users with failure flags or lost streaks when they miss a workout or eat off-plan. We wanted to build a supportive, adaptive companion that seamlessly recalibrates your schedule and pantry needs without any guilt.

What It Does:

Checkerboard Scheduler: Recalibrates your 7-day fitness routine automatically after missed days or diet changes.

AI Fitness & Meal Coach: Answers health, recipe, and exercise questions using an AI assistant.

Recipe & Exercise Search: Quickly surfaces targeted workout routines and meal ideas.

Receipt Scanner: Uses OCR to scan grocery receipts and track pantry items.

How We Built It:

Frontend: Built with Next.js (TypeScript) and styled using Tailwind CSS for a modern, clean UI.

Backend: Powered by a FastAPI (Python) server running Uvicorn.

AI & Search: Integrated ChromaDB for local vector search and vectorized context, paired with Ollama (Llama 3) for local LLM responses and Tesseract OCR for parsing receipts.

Made use of Gemini and OpenAI

Challenges We Ran Into:

Orchestrating state management between the Next.js frontend and FastAPI backend endpoints.

Managing database locks on vector store files during development and setup.

What We're Proud Of:

Delivering a full-stack, end-to-end working application with clean component architecture and smooth AI feature

Built With

• backend
• chromadb
• css
• database:
• fastapi
• frontend
• git
• llama
• llm
• next.js
• ocr
• ollama
• python
• react
• receipt
• tailwind
• tesseract
• tools
• typescript
• vector
• via