It was originally from a school group project, and it has made me think consider more about how can I make this complex idea turn live, with little prior experience in creating a functional prototype, ready to be converted into a real application.
It is basically a platform with different methods to detect stress (LSTM model to detect stress from typing cadence and scrolling patterns), allow students to build healthy de-stressing habits such as writing down the problems in a constructive method (not tested because of API constraints -- no tokens, but the functions are there). It would not save any of the private datas -- like what exactly you scroll and what exactly is typed. The focus is ALWAYS on PRIVACY.
Individually, I mainly used GPT 5.6 Terra from Medium for less demanding tasks to Extra High mode for more complex and clueless tasks.
There are some instances that the AI ran wild. Firstly, sometimes the idea it has come out overwrites the original idea. Secondly, the ideas that GPT 5.6 Terra came out with made me rethink about my original idea on its feasibility.
Completing this project is a feat. Only known this and started from Sunday 19 Jul GTC 1900 hrs!
AI is so prevalent, prominent and growing so unbelievably fast right now such that many have the chance to share their incredible ideas.
Hopefully, I can improve this project by having tokens to test out the functions and obtain more subjects within school to test out the usefulness of the application and spread out to other peers of different education level so that it would help them out :)
Log in or sign up for Devpost to join the conversation.