Try Meddy Today

Inspiration

Growing up with both parents working in hospitals and doctor's offices has kept me up to date with modern problems within the healthcare industry. One problem they always complained about was patient-hospital communication. When it came to applying treatments and prescribing medication, there was always a disconnect between the two. This often leads to unnecessary, avoidable errors that could sometimes even harm patients. I wanted to help connect patients with hospitals via an app.

  • One of our team members

Introducing Meddy, Your Healthcare Companion

Meddy bridges medical providers and patients together from day one to day done.

For the Patient

  • Finding a health provider is hard and intimidating. We made it more approachable with a tinder-style search
  • Your health data is always consolidated and available for your viewing and for easy first responders access
  • View your medications at any time and find the nearest pharmacy to pick them up at

For the Provider

  • Obvious medication conflicts are a dime a dozen when it comes to prescriptions. Meddy AI is here to read your patient's records for you and notify you of any dangerous combinations before sending the order off to the pharmacy
  • With the power of Meddy AI you can ask any medical questions about your patients concerning potential treatment options
  • View all your patients at a glance with our dashboard

How we built it

We built off the back of NextJS to create Meddy. Tailwind and ShadCN were essential to making everything look amazing in such a short time.

Our backend runs off Mongoose to reach our Mongo Database, and uses OpenAI's API to run Meddy AI. We also make a call to Google Maps' Embed API.

GenAI was instrumental to creating our dummy data for an app that feels realistic even in this early state.

Challenges we ran into

  • Deploying the site to Vercel was a real bear. We fought with ESLint so hard for hours over something that ended up being a logic error
  • Implementing Meddy AI was also a giant pain. OpenAI API documentation was not super helpful so we were mostly on our own
  • Styling issues constantly plagued the project. We kept on overwriting each other's files and bringing long-lost files back from the dead. We started off with feature branches, but quickly realized they were too cumbersome for the speed at which we needed to develop, favoring the good ol' "Can I push yet?" strategy.

Accomplishments that we're proud of

  • It's a real website! We did it, Mom!
  • Medical Provider Tinder was super fun to plan out and implement. It was one of the first feature-complete portions of the project
  • We got Meddy AI to work! And since we implemented it on the provider's side of things, we kept the usual ethical and security concerns to a minimum
  • It's a really, REALLY pretty site!!

What we learned

  • We should use Prettier from line 1 of code. Style guidelines are more essential than we thought
  • AI is super helpful for making official-looking dummy data and logos
  • Documentation can only get you so far. Sometimes you need to hit your head against the wall for a couple hours to find the solution
  • Constant communication about what we're doing and what our plans are kept us nice and AGILE during development. We avoided duplicating functionality and spread out the to do list items across ourselves

What's next for Meddy

  • Prettifying the provider view up to the standards of patient view
  • Proper database integration
  • Secure authorization for login
  • Patient data stored according to HIPPA guidelines
  • Bigger and better medication database
  • More fields and cross-play between database tables = more data for providers & Meddy AI
  • HIPPA-compliant messaging system within Meddy
  • Adding pretty animations and small user feedback and QoL changes
  • Providers using Meddy AI to modify their patient's records (add a new medication, treatment notes). no forms no hassle, just a conversation and it's in the database

Built With

+ 11 more
Share this project:

Updates