OptiHealth | Devpost

Inspiration

A common pattern was noticed among friends and family where they felt inspired to improve their physical well-being but were not sure where to begin when confronted with the multitude of information and advice online. Further, many had issues with the various generic workout splits posted online, with certain exercises being aggravating towards injuries and physical disabilities. Many also do not have the privilege of being able to afford a personal trainer for the long term. These issues and more prompted our idea of OptiHealth, an AI assistant that evaluates everyone's body on an individual level, providing a tailored workout plan that fits their specific needs.

What it does

OptiHealth, being a portmanteau of "Optimal" and "Health", is a web application that utilizes the Cohere API to generate a workout plan based on various criteria specified by the user. Optihealth is a website that uses AI to build a workout schedule based on the lifestyle of the user, and a prompt stating the types of accommodations they require. It aims to learn what the users' focus is when exercising and will use AI to determine what exercises they are capable of, to create an accurate workout routine. It was made with a focus on accessibility to support those who may suffer from chronic injuries or other physical limitations, and to make getting started easy for anyone.

How we built it

The frontend of OptiHealth was created with React.js, and backend with Flask. Each workout plan uses only one API call to Cohere, using strict prompting and regex to separate the information in an aesthetically pleasing and visually clean way.

Challenges we ran into

At first, the AI would oftentimes put an exercise deemed unreasonable in the workout routine (e.g: Telling someone with severe leg injuries to do squats), but adjusting the prompts for the AI to follow would later help solve this issue.

Accomplishments that we're proud of

This was the first time for all of us, to create a full-stack web application that utilized the API of an AI.

What we learned

We got a general introduction to how frameworks work through Flask, learned how the frontend communicates with the backend and how local hosting works.