FitFriend

FitFriend: An AI personal trainer that guides workouts, tracks progress, and listens to voice commands for hands-free, interactive coaching

Comment

Inspiration

We were inspired by a friend who goes to the gym very often. When he first started his fitness journey, he spoke often of his self-consciousness while performing exercises. He wished he could have a buddy to go with him, who could coach him through as he stepped out of his comfort-zone. We created FitFriend for others who are too scared to begin, and feel the same uncertainty.

What it does

FitFriend, allows users to have a fully functioning AI personal trainer. By speaking with this AI, it can recommend exercises, workout plans, and advice on anything gym related. Users can edit workouts suggested by the AI, and tailor it to their desired plan. Additionally, FitFriend tracks progress for users, allowing them to see how far they’ve come to reaching their goals.

How we built it

We used React for our UI components, and speech to text. Utilized Gemini AI to analyze context provided by the user and output a suitable exercise that fits their needs. Implemented ElevenLabs to bring our AI’s voice to life. Used MongoDB to track user fitness data and progress.

Challenges we ran into

Our biggest challenge was that our team consisted of two people, one of us being a beginner at hacking, and the other being a first year. With half the man power and inexperience we took a while to get in our groove and find our roles in building this project.

Accomplishments that we're proud of

We were able to create a webapp with lots a features, using Elevenlabs, Gemini, MongoDB.

What we learned

Using a diverse TechStack, we improved our understanding of online sites, and their utilization of many different types of tools. Learned how webapps function, and how they interact with backend.

What's next for FitFriend

The end goal for FitFriend is to allow the user to communicate with the AI coach, almost as if they were right in front of them. This means: Setting up a mobile version, which is more practical for uses in the gym. Adding quality-of-life changes. A simple feature like a hot word would make conversations more smooth, and flesh out the app. Creating our own actual LLM. For now we use gemini, but we would like more detail, better recommending, and greater interaction with the user.

Built With

Share this project:

Updates