Inspiration

AEGIS was inspired by the increasing impact of natural disasters and the challenges emergency services face when coordinating resources. We wanted to explore how AI and resilient communication could help emergency workers respond faster, allocate resources more effectively, and ultimately help protect more lives.

What it does

AEGIS is an emergency coordination platform that uses AI to analyse incidents, determine the resources required, and support emergency response. Gemini processes emergency reports and structures the information, while ElevenLabs turns the resulting response into a clear voice announcement for emergency workers.

How we built it

We built AEGIS using React and TypeScript for the interactive frontend, with CSS for the visual design and animations. The backend was built in Python using FastAPI, with Gemini for AI-powered incident analysis and ElevenLabs for text-to-speech emergency broadcasts.

Challenges we ran into

We found that the Gemini models we were originally using were not responding to our API requests. After trying multiple models and developing debugging tools to find where the issues were we found a model, Gemini 3.5 Flash Lite, which was reliable. Another was designing a frontend that made a complex distributed system immediately understandable. Had to manage multiple pieces of React state so network failures, node isolation, incident injection and reconciliation all updated consistently. Building the interactive network visualisation and keeping the UI responsive while maintaining the emergency command-centre aesthetic required a lot of iteration.

Accomplishments that we're proud of

We are particularly proud of successfully integrating Gemini and ElevenLabs to turn emergency reports into AI-analysed, spoken alerts. We are also proud of the dynamic dashboard, which makes incidents, network failures, autonomous responses and recovery easy to visualise and interact with.

What we learned

We learned the importance of attention to detail, as many of the issues we encountered came down to small syntax, configuration or integration errors that could take time to identify. We also learned how important it is to connect different technologies reliably when building a system with multiple moving parts.

What's next for AEGIS Symphony

We want to develop AEGIS into a more dynamic resource-allocation system by giving it real-time knowledge of available emergency resources. For example, if a new higher-severity incident occurs, AEGIS could reassess existing allocations and recommend redirecting suitable resources to the more urgent incident, with a new voice announcement communicating the change. We would also improve the operational map to provide a more accurate representation of the city and allow responders, infrastructure, incidents and communication failures to be visualised in greater geographical detail.

Built With

Share this project:

Updates

Submission history