Aurora Assistance

When in need of medical assistance, but it's not sever enough for immediate emergency assitance, but you also prefer a doc to check you out over a simple 811 call.

Comment

Inspiration

This idea is strongly inspired from the sponsor challenge from Dialogue. We did our best to implement a an application that would be as realistic as possible.

What it does

Aurora Assistant is an AI assistant that helps assess non-emergency medical situation in real time. It analyzes symptoms, provides guidance and can automatically book clinic appointments on the user's behalf. By handling calls and scheduling using Text-to-Speech technology, it reduces stress and saves times when quick medical action is needed without unnecessary panic.

How we built it

MERN Stack + Gemini API + Tailwind CSS. + ElevenLabs Batch Calling API

Challenges we ran into

Data consistency and enforcing AI constraints. Implementing Google OAuth and figuring out a way to implement ElevenLabs outbound batch calling to a real phone number.

Accomplishments that we're proud of

Making the phone call AI actually work. Having a finished MVP

What we learned

  • How to integrate an LLM into a project and giving it a voice.
  • How to login user via Google OAuth and storing user data into a mongoDB.
  • How to implement ElevenLabs outbound batch calling to a real phone number.

What's next for Aurora Assistance

Fine-tuning the chatbot to provide more accurate assessment to users. Make assistance available without internet connection Implement extra features such as multiple profiles

Video link: https://drive.google.com/drive/folders/14fyQVHfI34SzkH1c3G8lsWXImrqgOHX5?usp=sharing

Built With

Share this project:

Updates