FridgeFit

FridgeFit scans your fridge photo to instantly create personalized, goal-based meal plans using ingredients you already have. Eat smarter, waste less, and hit your fitness goals.

Comment

About the Project – FridgeFit

Inspiration

We built FridgeFit to solve a common problem: figuring out what to cook with the food you already have. Too often, ingredients go to waste or people struggle to eat healthy due to lack of time, planning, or knowledge. Inspired by fitness tracking, AI innovation, and the frustration of aimless snacking, we created FridgeFit to turn your fridge into a personalized nutrition coach.

What it does

FridgeFit uses AI to generate goal-based meal plans directly from a photo of your fridge. Users can:

  • Upload a photo of their fridge contents
  • Receive 3 customized meals based on visible ingredients
  • Tailor plans to fitness goals (e.g., bulking, cutting, clean eating)
  • See macros for each meal: calories, protein, carbs, fat
  • Get recipes that match their preferences and dietary needs
  • Waste less food by cooking with what they already have

It’s like having a personal nutritionist and chef, right in your pocket.

How we built it

FridgeFit is built using:

  • React – For a smooth, responsive frontend
  • Supabase – Handles user authentication and stores fitness profiles
  • GPT-4o with Vision – Powers the ingredient recognition and meal generation
  • Edge Functions – Process the fridge image and generate structured meal plans
  • Tailwind CSS – For clean, responsive UI across mobile and desktop
  • CORS & Error Handling – Ensures secure and smooth backend communication

Challenges we ran into

  • Accurately identifying ingredients from diverse fridge images
  • Designing a strict, readable JSON format for meal generation
  • Adapting AI output to match real-world user preferences
  • Generating macros that are consistent and goal-aligned
  • Ensuring low latency even with image processing and AI calls

Accomplishments that we're proud of

  • Created an end-to-end meal generator from photo to macros
  • Developed a strict prompt system that guides GPT-4o to output valid, useful results
  • Successfully used Vision AI to interpret real-world fridge inputs
  • Built a fitness-friendly app that’s both practical and fun
  • Deployed as a clean, mobile-first web app with real-time feedback

What we learned

  • How to use GPT-4o Vision for structured reasoning tasks
  • Prompt engineering for image-to-text and logic conversion
  • Structuring AI responses for frontend rendering
  • Importance of dietary personalization in fitness tech
  • Full-stack deployment and debugging of AI-driven apps

What's next for FridgeFit

  • Launch daily fridge scan streaks to keep users consistent
  • Add barcode scanning and manual ingredient entry options
  • Let users favorite and save meal plans
  • Build a grocery list generator for missing ingredients
  • Offer premium coaching tips based on nutrition trends
  • Partner with fitness influencers and nutritionists for viral growth

FridgeFit makes healthy eating effortless—personalized, goal-driven meals using what you already own, all powered by AI.

Built With

Share this project:

Updates