Inspiration

Following many natural disasters, we heard news about first responders being overwhelmed and cannot work at full efficiency. The idea of triage, in our opinion, could help this situation – knowing the urgency of different emergencies, we are able to allocate resources to the maximum efficiency. We believe this app is very rich in social goods and could help first responders in times of crisis.

What it does

In short, Zeroth Responder is an AI 911 agent. Zeroth responder, as its name suggests, is used before the first responders take a call. It listens to the descriptions of the caller to judge their circumstances, asks follow-up questions, and effectively note down and classify the situation based on its urgency. In addition, it allows first-responders to have first-hand information about the callers in summarized forms and prioritize their resource allocations.

How we built it

We stored a 911 training manual into a vector database powered by Milvus, then queried it in conjunction with LLMs to create responses to the caller and assess the situation.

Challenges we ran into

We had much trouble setting up connection with the database, as well as making various parts of our stack integrate with each other.

Accomplishments that we're proud of

We are proud that the app is functionally complete and could contribute to society.

What's next for Zeroth Responder

We are integrating with supabase to create a dashboard for first responders to assess the severity of each call and to deal with them appropriately.

Built With

  • nextjs
  • openai
  • supabase
  • vercel
  • zilliz
Share this project:

Updates