DocuMed

Create Report
Dashboard
User Profile

Inspiration

We built DocuMed because we grew up in a community where high medical bills often meant skipping hospital visits. Recognizing that billions face similar barriers—whether due to cost, location, or access—we created DocuMed to make healthcare accessible for everyone, no matter the circumstance.

What it does

Account Creation / Login: The user starts by creating an account or logging in.

Dashboard Access: Once logged in, they are taken to their dashboard, displaying all previous medical reports. Here, they can choose to view or delete any report.

Profile Management: The user can update their profile information, which helps us provide tailored recommendations based on their unique health needs.

New Report Creation: Most importantly, users can generate a new report by speaking with our hyper-trained AI assistant, enabling them to receive immediate, personalized insights and recommendations.

How we built it

We built this app on a tech stack of TypeScript on a React.js library. Our API calls were handled by axios., and our backend was python built on Flask. In order to encorporate our AI-features. We created a Pinecone Vector Database, a supabase database to handle user/report information, and used langchain.

Challenges we ran into

Throughout development, we encountered several key challenges. For example, handling retrieval-augmented generation queries was difficult due to limited documentation. Our biggest challenge, however, was implementing speech-to-text and text-to-speech functionalities. Managing the various libraries and files required for this feature involved hours of trial and error to achieve efficient performance

Accomplishments that we're proud of

Our report generation feature is what we’re most proud of. It creates reports in real time for optimal speed, using multi-threading to boost both accurate responses and efficiency. With the extensive AI integration, maintaining a fast, responsive app was a priority, and we're especially proud of how quickly it performs. This keeps users and engaged and makes it a friendly app for all ages, and all countires.

What we learned

We learned so much about foundational software engineering concepts such as multi-threading, concurrency, mutex locks. live-time processing. These were foundational building-blocks for creating our app. Not only that, but we also learned how to manage/tackle problems as a team - we had to split up work, and also organize it based on our own experiences which we were able to do well!

What's next for DocuMed

We have 2 very foundation concepts we want to integrate into DocuMed next

Doctor Accounts: We started building this but ran out of time, essentially we want users to be able to login to this app as a doctor and connect with patients by linking them to that same report. We also want to enable messaging in the app so their is no boundary where you can connect with medical preofessionals.
Language Translation: Currently our app does not support our AI-nurse speaking in multiple languages (although a user can speak in any language they want). We want to change this so that their is no barrier for any person.