AI Dietary Planner

Inspiration

We all know that what we eat directly affects how we feel, how much energy we have, and how well our bodies function. But with so much conflicting advice on the internet, figuring out what’s actually good for you can be overwhelming. That’s where our AI Dietary Planner comes in.

Our goal is simple: to make healthy eating easier and more personalized for everyone. We believe that good nutrition isn’t one-size-fits-all. Whether you’re managing a medical condition, working toward a fitness goal, or just trying to eat better, the right plan can make all the difference.

What it does

By analyzing your unique needs such as age, health conditions, and personal preferences, our AI provides clear, personalized dietary guidance using the OpenAI API to keep you on track and improve your health. Simply input your information, and you’ll receive customized meal ideas and nutrition advice aligned with your goals.

How we built it

We built this application using React.js and TypeScript for a fast, reliable frontend, with Tailwind CSS and Bootstrap for a clean, responsive design. On the backend, we used FastAPI to handle requests smoothly and Node.js to manage dependencies. This combination allowed us to create a seamless and user-friendly dietary planning experience.

Challenges we ran into

Initially, we planned to fine-tune an open-source large language model for our use case. However, we quickly realized that training the model was too computationally expensive, even with GPU resources, and achieving the desired accuracy was unrealistic within the limited time we had. Pivoting to our current project midway through the hackathon was challenging, especially since this was our first hackathon and our first full-stack application, however, with perseverance, we were able to have a functioning application in the end.

We also faced difficulties with Git, as it was our first time using it for collaborative development. Fortunately, with guidance from our mentors, we were able to resolve the conflicts and keep the project on track.

Accomplishments that we're proud of

We're proud that we finished the project despite the many challenges we came across. At one point, we all had felt helpless, almost close to completely giving up on this project. However, we are beyond happy to not only create a full-stack application but to be able to help individuals stay on track with their diet within the allotted time.

What we learned

We learned many things throughout the process that we will all take away and apply in other aspects of our lives. Soft skills: collaboration with teammates, effective idea communication, perseverance through challenging errors. Hard skills: full-stack development workflow, Git, API creation, prompt engineering.

What's next for AI Dietary Planner

Moving forward, this application could be expanded by adding more advanced features such as tracking user progress as they work towards their health goals or perhaps building a community where users can be able to share their experiences and meal plans. We could also implement model adaption and enhancement techniques such as fine-tuning or Retrieval-Augmented Generation (RAG) to further improve the generation of dietary plans for users.

Built With

• fastapi
• github
• node.js
• openai
• python
• react.js
• tailwind
• typescript
• vscode