Crave Well

main page
Landing page
meal analysis and recommendation
firebase

Inspiration

Imagine walking into a restaurant, craving something delicious, only to be hit with the weight of dietary restrictions. Whether it’s high cholesterol, diabetes, muscle-building goals, or religious/ethical diets—your options shrink, and joy disappears from the experience of food.

We were inspired to solve this frustration.
CraveWell was born from a simple belief: your diet should empower you, not limit you.

What it does

CraveWell is an AI-powered web app that helps users make informed, personalized food decisions without sacrificing flavor or freedom.

Users upload a photo of their meal
Our system detects the food items, ingredients, and nutrition data
It then provides a personalized health analysis based on the user’s dietary conditions, body metrics, activity level, and main health goals

Whether you’re trying to gain muscle, improve heart health, manage diabetes, or just feel good—CraveWell gives you real-time feedback and recommendations for your unique body and lifestyle.

How we built it

We combined several powerful tools and models:

Google Cloud Vision API to identify food items from images
Cohere (Command R+) for generating personalized health recommendations via smart prompts
USDA API to retrieve rich nutrient data for each ingredient
Firebase to store user profiles, conditions, goals, and meal data
Streamlit and VS Code for local development and UI integration

We worked in sync across frontend, backend, and AI prompt engineering to make this work smoothly.

Challenges we ran into

Structuring ingredient-level USDA nutrition data into meaningful summaries for full meals
Building prompts that adapt to various health conditions and fitness goals while remaining accurate and actionable
Handling inconsistent image recognition data and filtering out non-edible or generic labels
Managing different units and formats for nutrient values across APIs
Making the system feel personal, not generic

Accomplishments that we're proud of

We built a working AI pipeline from image → food → nutrition → personalized health advice
Created an experience that is inclusive of many diets: vegetarian, diabetic, halal, low-sodium, fitness-oriented, and more
Developed dynamic prompts that change based on real user data (height, weight, gender, condition, goal, activity level!)
Designed the foundation for a tool that could make a real difference in people's lives

What we learned

How to build an AI product with real impact and empathy
How to chain APIs (Vision + USDA + Cohere) and align them with user context
How important prompt engineering is for healthcare-based AI systems
That nutrition is deeply personal—and the best product is one that listens first, and speaks second

What's next for CraveWell

Live user testing with people managing real dietary conditions
Expanded dietary profiles, including allergies, religious restrictions, and performance-focused macros
Meal suggestions that align with user goals
Firebase login & profile dashboard
Multilingual UI to expand access

We're just getting started. With CraveWell, we're redefining what it means to eat freely and live fully—with zero compromise.