Santé

LOGO

Inspiration

Sante started as just an idea. A mere conversation among our friends. It became a heavily discussed topic when the news about flo came out. The news enraged us but at the same time , what can we do? What could we have done? About apps that sold our data as if they were selling mere candies. Apps that succumbed to propagandas concerning women’s health. If not just that, there were other problems too. Health apps didn't even consider energy levels,circumstances a person is in or even the mood. They had rigid weekly or monthly plans about diet ,exercise etc. Most apps assumed that every woman had a perfect 28 days menstrual cycle. The list goes on. We really wanted an app that deeply cared about women and their health issues.The opportunity presented itself in the form of this hackathon and our idea had the chance to shine. So this became our main inspiration and the force behind all of our hardwork. This was the beginning of Sante. So without further ado, let me introduce you to our app that we have brought together for you

What it does

Sante begins asking four questions:energy,discomfort,mood and sensory load. Very simple but very crucial questions for your day.it keeps your goal , just changes the way you approach it and makes it easier to complete. It asks six simple questions to start. • How much energy do you have today? • How comfortable does your body feel? • Where is your mood today? • How much sensory load are you carrying? • Anything today that needs care?

The user can answer the first 4 questions on a scale of 1-5 or can also use the option that he/she/they aren’t sure how they feel. This makes it less stressful. The last question takes into consideration any additional changes that need to be taken care of. Taking into consideration your capacity for today it creates a plan for you that adapts to your energy levels. Sante also explains the changes made to fit your needs so users know why they are doing what they are doing. It also keeps record of what you did last time and your feedback about the day. It also provides a 20 second check-in that users can do in-between their day if the mood changes or something happens. It takes into consideration that all days don’t go as we plan so user is free to make changes during any part of the day

How we built it

Santé is built on Next.js and vercel with a supabase database keeping the data safe. So when a user checks in , their entry pass through the code first not the AI and when its passing through the code , the code runs a quick safety check and works out what the plan should look like even before AI is ever involved. after that our AI aka Luna steps in and shape the plan it only chooses the movement we already trusts so it can never shape something unsafe for user's body. Every information is locked to the user alone where important safety decisions is made by our own code. Even if AI failed Sante never. theres backup and still gives user its safe space with or i would say without the help of AI.

Challenges we ran into

One of the main Challenges was deciding what AI was even useful and where and how much authority it should have because lettign a model choose safety limits is not so reliable for the system and the users so we had to separate the determinstic safety and constraints from flexible work. Another challenge was designing for someone having low capacity putting ourselves in their shoes,screen can look all polished but for someone whos overloaded it still feels overwhelming. Another challenge was many bugs appeared when we were in production journey, like RLS protecting Rows not columns , blocked verdict returning a usable plan , rate limit not covering the unauthenticated users, calm mode being desync, React hydration errors, blank greeting on the red-flag path ,indentity clean-up deleting a check-in and then we had AI behavioural issues .

Accomplishments that we're proud of

We can honestly say our safety relies on the code not on the model and we proved it by atttacking our own app from outside with public key. Sante genuinely adapts and most importantly it does not push any women to same standards.

What we learned

The most important part was that AI is only trustworthy if you know what you are building around it . Our most important safety would never touch the AI at all . we also learned to review bugs , codes and many more things at such short time .

What's next for Santé

In the future we will add the app in IOS and andriod. And will linkage with samsung health and apple health to import health data because it could be update to tracker app too .