NutriSnap

Hero section
Setup
Home
BMR

Inspiration

In this busy world, individuals have difficulty staying on a balanced diet, particularly in areas where they cannot find cheap nutritionists. Most households, especially semi-urban and rural ones, unknowingly eat unbalanced diets that cause health problems such as diabetes, obesity, and malnutrition. This led us to build NutriSnap AI — a user-friendly, AI-driven smartphone app that allows anyone to eat intelligently with a simple photo of what they eat.

What it does

NutriSnap AI allows you to snap a pic of what you eat or upload a pic, and the app instantly identifies food items, guesses calorie counts, and provides a full analysis of both macronutrients (carb, protein, fat) and micronutrients (vitamins, sodium, calcium, iron, etc.).

They can also enter their height, weight, age, and sex so they receive a custom daily calorie range. The program monitors their meals, keeps a history of daily activity, and shows history of food data on a dynamic timeline view.

It is what makes NutriSnap AI special that it has a "Smart Swap" feature — when a user is suffering from a condition such as diabetes or high blood pressure, the program gives healthier advice than just alerts. For instance, instead of saying "avoid fried rice," it would "try brown rice with sautéed vegetables."

How we built it

We developed NutriSnap AI with a mix of Firebase for user authentication and database services and OpenAI API for the identification of food items and prediction of nutrients. The UI was made simple, easy-to-use, and health-oriented — with a soothing green theme that exudes nutrition and health.

Problems faced by us

It was also one of the biggest challenges obtaining the AI estimate of the nutrients accurately, especially for highly complex Indian recipes with a lot of ingredients. We had some small integration issues when connecting the photo-recognition API to Firebase user tracking, also. Keeping the design simple, mobile-first without getting too "tech-heavy" was another challenge, all within the 48-hour hackathon time constraint.

Accomplishments that we're proud of

We are happy to have built a working project.

It does not merely track calorie counts but gives one a sense of insight, healthy living awareness, and individual nutrition recommendations through AI. That it recognizes Indian food names and makes region-specific recommendations is something on which we are particularly proud.

What we learned

We learned how to effectively unify AI-powered image recognition, integration with databases, and custom user experiences under a single application. The hackathon educated us about the need for teaming, modular coding, and blending innovativeness with usability.

What were you constructing for?

Our task was to challenge ourselves for creating a solution between tech and healthy living.

What broke?

We experienced problems that ranged from the AI model misreading multi-item orders and incorrect calorie displays when presented for testing. Data syncing between Firebase and the front-end was also lost when changing users at times.

Why does it still somehow work?

Even with a few minor glitches, the AI on NutriSnap delivered on its primary functions — identifying food, estimating calorie counts, providing detailed nutrient facts, and suggesting healthier alternatives. Lacking extensive training data and time, the software showed that user-friendly AI-driven nutrition analysis is within reach and effective.

Survey Link:- https://docs.google.com/forms/d/e/1FAIpQLScRafXPPLzQ4b45zzBHEt4zKRZru4ZmlbT1NsCqgwnJJGjpXA/viewform?usp=publish-editor