GauchoCrub

GauchoGrub — AI-Powered Meal Planning for UCSB About the Project

GauchoGrub is a web application that creates personalized meal plans for UCSB students using the university’s live dining hall menu. Students enter information such as height, weight, dietary restrictions, and health goals, and our AI generates a realistic daily meal plan using only foods that are actually available on campus that day. Our goal is to make healthy eating at UCSB simpler, smarter, and more accessible.

Inspiration

Like many UCSB students, our team has experienced the daily struggle of deciding what to eat at the dining halls. Menus change constantly, nutrition information can be overwhelming, and students with dietary restrictions or fitness goals often feel lost. We wanted to solve a real campus problem with technology we were excited to explore — AI-powered personalization. GauchoGrub was born from the idea: what if choosing healthy meals on campus took seconds instead of guesswork?

How We Built It

We built GauchoGrub as a web application with a simple but effective pipeline:

We fetch the live UCSB dining hall menu from the public UCSB dining site.

The user enters personal data such as dietary restrictions and wellness goals.

We send both the menu data and user profile to the Gemini API through OpenRouter.

The AI returns a customized meal plan using only available menu items.

The plan is displayed on the website in a clean, easy-to-read format.

By keeping the interface simple and focusing on core functionality, we ensured the project was fully demoable within hackathon time constraints.

Challenges

One of our biggest challenges was working with real-world data. The UCSB dining menu is not formatted for easy parsing, so we had to experiment with extracting usable text for the AI prompt. Another challenge was scoping — we had many ambitious ideas like nutrition tracking and account systems, but we learned quickly that finishing a polished MVP was more important than overengineering. We also encountered environment setup issues while configuring Python and Flask on macOS, which taught us valuable debugging and dependency management skills.

What We Learned

This project taught us how to:

Integrate external web data into an application workflow

Work with LLM APIs for structured, goal-oriented generation

Build a full-stack web prototype under time pressure

Scope a project realistically for a 24-hour hackathon

Most importantly, we learned how to turn a real user problem into a functional product prototype.

Closing Thoughts

GauchoGrub demonstrates how AI can make everyday student life more convenient and healthier. While this is just a prototype, the concept could be expanded with nutrition tracking, saved profiles, and long-term meal planning. We’re proud of what we built — a working, practical AI tool designed specifically for our campus community.

Demo: https://drive.google.com/file/d/1Znipz5yOgwl5zDeZfOHEKfRTgXY6Wl83/view?usp=sharing