Fridget

• 

  Sign Up page where users can submit their dietary preferences

• 

  Login Page

• 

  Homepage (Recipe Recommendation page) where recipes are ranked by their current ingredients inventory and dietary preferences

• 

  Detailed information of a recipe with description, ingredients, steps

• 

  Ingredients page where users can add/remove their ingredients in their fridge

✨Inspiration

We all know the struggle of staring into the fridge, wondering what to cook with random leftover ingredients.

Wasting food is a problem, and we wanted to create a fun, smart, and efficient way to turn what you already have into delicious meals.

Fridget is all about making the most out of what’s in your fridge while minimizing waste and ensuring easy access to delicious, practical recipes for everyone with AI!

💪What it does

Fridget is an AI-powered recipe generator that takes the ingredients you already have and suggests customized, easy-to-follow recipes.

Given the dietary information such as dietary preferences, allergies, and spice levels of a logged in user, it recommends and ranks recipes that the can make use of with what they have in the refrigerator.

Along with the given recipes, it identifies missing ingredients and notifies the user.

🚜How we built it

🚧Frontend: React, Chakra UI, Material UI
🚧Backend: Spring Boot, MongoDB, and Redis for efficient data storage and fast processing (Access Token and Refresh Token)
🚧AI & Machine Learning: OpenAI for recipe generation, Scikit-learn & SpaCy for ranking and ingredient matching
🚧Authentication & Security: JSON Web Tokens (JWT) for secure login and user data protection.
🚧Python Integration: Seamless interaction between Java (Spring Boot) and Python scripts for smart recipe sorting and recommendations.

Challenges we ran into

🏃🏻‍♀️running Python file in Java (since it was our first time)

🏃🏻‍♀️AI model was highly sensitive to rewards and penalties, making it difficult to find the optimal parameters

🏃🏻‍♀️creating unwanted dependencies/infinite loops through nested hooks(React)

🏃🏻‍♀️utilizing Github to collaborate when more than one person modifies a single file

🏃🏻‍♀️dealing with web scraping dynamically to generate an image for each recipe

Accomplishments that we're proud of

😎rendering image using google search API on the live run

😎successfully completed the development of an AI-based system that is user-friendly

😎completed the implementation of an MVP in a limited time

😎integrating different technologies together despite not knowing how to use other ones

What we learned

✏️how to use multiple hooks inside multiple UseEffects

✏️that AI can be utilized even for small tasks to enhance human convenience

✏️reminded that just because the code runs on one computer doesn't mean it'll run on another

✏️building the system architecture and communicating with others is as important as the project result itself

What's next for Fridget

🚀 Personalized meal planning based on user preferences and past cooking history.

🚀 Community recipe sharing, where users can upload and rate fridge-friendly recipes.

🚀 Smart shopping list suggestions, so users know exactly what to buy based on missing ingredients.

🚀 Integration with smart fridges to auto-detect available ingredients and generate recipes instantly.

🚀 More advanced AI models for better recipe optimization and cooking instructions.

Built With

• chakra
• java
• jwt
• materialui
• mongodb
• openai
• python
• react
• redis
• scikit-learn
• spacy
• springboot