FridgeMind AI

• 

  An AI-powered suggestion of what to eat, based on the food that is expiring food

• 

  An empty fridge

• 

  A fridge full of food: some expiring soon, some expiring later

Inspiration

People often don’t know what to eat first. Food left in the fridge expires and gets thrown away. This leads to wasted money and unnecessary environmental impact. I wanted to build something simple that doesn’t force people to change their habits but instead offers many options to help them use what they already have, just at the right time.

What it does

FridgeMind AI helps users decide what to eat based on what’s in their fridge and when it expires. Users can: -add products with expiration dates -see which items are urgent -get instant AI-powered suggestions on what to eat -see approximately how much CO₂ they can save The focus is not on recipes, but on making the right decision at the right time.

How I built it

I used Python and Streamlit for the interface and website hosting, the Groq API (LLM) to generate food decisions, and simple local storage (JSON file) to keep track of products in the fridge. The system sorts items by expiration and sends that data to the AI, which generates suggestions and recipes.

Challenges I ran into

Making AI outputs short, useful, and not generic. It was especially important to ensure the suggestions actually make sense - real, tasty combinations, not random mixes like fish, chocolate, and milk. Handling deployment challenges, including API keys and Streamlit Cloud. At one point, I accidentally exposed my API key, which led to it being revoked and forced me to rethink how to handle secrets securely.

Accomplishments that I'm proud of

Turning a simple idea into a working AI-powered product. Building a system that focuses on decisions, not just suggestions. Successfully deploying and making it usable for others. Creating something that connects everyday behavior with real-world impact.

What I learned

How to integrate LLM APIs into real applications (and make them private). How to work with AI and write a successful prompt. How to debug and deploy a project under time pressure (was working until 2 am).

What's next for FridgeMind AI

Add image recognition to detect food and its expiration date automatically. Improve personalization based on user habits. Add smarter impact tracking (CO₂, waste reduction). Turn it into a fully working mobile app.

Built With

• python
• streamlit