• 

  Homepage

NutriAI - AI-Powered Meal Plan Generator

Inspiration

The inspiration for NutriAI came from the idea of making healthier eating habits accessible to everyone. We wanted to leverage the power of artificial intelligence to create a tool that simplifies meal planning, tailored to individual preferences and health goals. The project was envisioned as part of HackHive 2025, showcasing how AI can be used to solve everyday challenges.

What it does

NutriAI is a web application that generates personalized meal plans based on user dietary preferences and calorie goals. By integrating AI into meal planning, NutriAI aims to make healthy eating more convenient and approachable for all.

How we built it

NutriAI consists of two main parts:

• Backend: Built using Node.js and Express, the backend leverages Hugging Face's microsoft/Phi-3.5-mini-instruct model to generate meal plans. The backend processes user input, constructs a prompt for the model, and parses the generated response to return a clean and actionable meal plan.
• Frontend: Built using React.js, the frontend provides a clean and interactive interface for users to input their preferences and view their personalized meal plans.

The project was developed in just one day as part of HackHive 2025.

Challenges we ran into

One of the biggest challenges we faced was finding a suitable AI model for our application. We explored several models, some of which were either gated, too large for our requirements, or did not produce the desired outputs. After extensive testing, we found the microsoft/Phi-3.5-mini-instruct model to be the best fit for our needs.

Additionally, integrating the model with the backend API and handling issues such as CORS policies and proper tokenization required a lot of debugging and problem-solving.

Accomplishments that we're proud of

We are incredibly proud of:

• Developing a fully functional AI-powered meal planning application in just one day.
• Successfully integrating Hugging Face's model into the backend.
• Creating a clean and user-friendly frontend interface for the application.
• Overcoming significant challenges in model selection and debugging to deliver a working solution.

What we learned

This project was a crash course in AI model integration and deployment. We learned:

• How to interact with Hugging Face's API and test various models.
• The importance of constructing clear and concise prompts to get meaningful outputs from AI models.
• Best practices for building and debugging full-stack applications in a short timeframe.

What's next for NutriAI

Looking forward, we plan to:

• Implement an account system so users can save and revisit their preferences and meal plans.
• Add more AI-powered features, such as generating shopping lists based on meal plans.
• Expand the frontend to include detailed nutritional breakdowns and recipe suggestions.
• Explore more advanced AI models for greater personalization and accuracy in meal planning.

Built With

• express.js
• node.js
• react.js