Nutr.IA

About the project

Nutr.IA is an open-source web app that uses artificial intelligence to generate meal plan suggestions. Generation takes into account the physical data, goals and activity level provided by the user.

Problem

Generic meal suggestions do not account for each person’s data and goals. The project explores how to organize that data into a workflow for generating, reviewing and storing meal plans.

Solution

Users enter their details and goal, and the application generates a suggested plan with meals, times and foods. Results are saved in a history where they can be reviewed, exported as PDF or deleted.

My role

I developed the product flow, AI model integration, authentication, plan storage, PDF export and interface. I also made the project available as open source.

Key features

• Form with name, age, weight, height, sex, goal and activity level
• Suggested meal plans with meals, times and foods
• History of generated plans
• Detailed plan view
• PDF export
• Email and password or Google sign-in

Architecture and technical decisions

The application uses Next.js, TypeScript and Tailwind CSS. NextAuth handles credential and Google authentication, while Firebase stores each user’s plans.

Google Gemini receives the submitted information and generates the plan content. Its response is organized into a meal structure used by both the interface and PDF export. Registration data is validated with Zod before being sent to the API.

Outcome

Deployed, with source code available on GitHub.

Technologies

Next.js, TypeScript, Google Gemini, Firebase, NextAuth, Tailwind CSS, Zod and jsPDF.