CyberFood

• 

  Final Design

• 

  Final Design

• 

  Initial Mockup

• 

  Initial Mockup

• 

  Prototyping

• 

  Prototyping

Elevator Pitch

In a world where food our food production takes place further from home, using industrial processes, it's more important than ever for people to understand the ingredients that make up items in their pantry. In the future we might be eating nutrition cubes but before it's too late it's our job to educate the public on how processed their foods are. CyberFood swiftly explains how processed the ingredients are and scores the food according to the NOVA Classification System using artificial intelligence.

Inspiration

In a cyberpunk future filled with megacorperations, fresh food and produce might cease to exist, only to be replaced with ultra-processed foods with impossible to decipher ingredients lists. We felt it fitting that CyberFood take on the herculean task of warning members of the public before it's too late!

What it does

Cyberfood uses OpenAI's GPT-4o to identify food items and queries the Food Data Central API to fetch the ingredients list. Once it gets the ingredients list, the model classifies the ingredients into NOVA food classifications based on how processed the ingredients are. Once the model has done it’s classification, we compute a score of how “bad” the food is based on how many of its ingredients are processed and to what degree. The user is then shown a score and a list of the ingredients to better inform them when it comes to their food choices.

How we built it

• Mockup: Figma
• Backend: Python + Flask
• Frontend: Typescript + React + Vite

Separated Tasks

• React components
• API endpoints

Challenges we ran into

• Flask learning curve
• React routers
• Using Tailwind for the first time

Accomplishments that we're proud of

• Organized code on GitHub
• Our prompts work well

What we learned

Learned a lot about the nitty gritty details of Flask and also the OpenAI API.

What's next for CyberFood

• AI voice (e.g., Duke Nukem, Solid Snake) to speak to the user and maybe explain why the decision was made for each of the ingredients.

Built With

• flask
• nix
• openai
• pydantic
• pytest
• python
• react
• typescript
• vite