FitFreak

FitFreak, an platform to gamble your fitness with other fitness freaks. Users pay eth to join fitness competetions, motions tracked in realtime using AI, and the winner gets instant blockchain reward.

Comment

Inspiration

Actually i have seen those application like dream 11,, but that seems like gambling,, and more friend who is a gym guy but he doesnt make much money, so yeah the midnight conversation with my gym friend inspired me to build this application ## What it does Basically we can gamble out fitness to earn, If am a fitness freak person and i want to participate in competition and get instant rewards, this platform would help a lot, simply the person pays certain amount of eth to participate in the competition, and some other users would do the same thing,, and they would perform an live workout event and their workout would be monitored by AI in realtime, if an user won that competition then he would get reward(Coin/Token) instantly. And if you are not a gym freak, no worries.This platform provides the user to bet on other players, and can also watch those competitions in live ## How we built it The frontend is build with react+Vite, And for ai Mediapipe and opencv and backend with flask , blockchain contract written in solidity.

Challenges we ran into

One of the main challenges was handling realtime motion tracking with consistent accuracy across different lighting conditions and camera qualities. Integrating mediapipe with the backend and syncing AI results with blockchain events was also tricky. Another challenge was creating a smooth live competition flow without latency issues. Handling crypto payments and making sure the transactions triggered rewards instantly took time to figure out.

Accomplishments that we're proud of

We are proud that we were able to build a working prototype that tracks workout motions in realtime, connects users in a fitness competition, and rewards winners instantly through blockchain. Getting AI tracking + live events + blockchain together in one platform itself is a big accomplishment.

What we learned

We learned a lot about mediapipe pose detection, realtime video processing, solidity contracts, and integrating blockchain with a traditional backend. We also understood how to optimize AI models for live use and how important latency management is for realtime competitions.

What's next for FitFreak

Next, we are planning to improve the accuracy of the workout detection, add more competition types, introduce social features, and expand the blockchain reward system. We also want to make the platform available on mobile and build a more scalable tournament system.

Built With

Share this project:

Updates