BlazeGuard | Devpost

What it does

BlazeGuard is a voice-powered medical assistant that helps users access their stored medical information and provides real-time emergency guidance. Users can talk to the AI agent, which retrieves personal medical details by name and delivers emergency response instructions using Retrieval-Augmented Generation (RAG). How we built it:

Frontend: Built with Next.js and React for a responsive and intuitive web application.
Voice AI Agent: Developed using Python, integrating LiveKit for real-time voice communication.
AI Models: Utilized Gemini-2.0-Flash-Exp for agent reasoning and Gemini TTS for natural-sounding speech output.

Challenges we ran into

Ensuring fast and reliable voice interactions with minimal latency. Structuring medical data retrieval in a way that is both efficient and secure. Getting the Gemini API working with LiveKit.

Accomplishments that we're proud of

Successfully integrating real-time voice communication with AI-powered medical assistance. Building a functional and responsive web app in a short time frame. Implementing a structured RAG system for accurate emergency information retrieval.

What we learned

Optimizing voice AI interactions for low-latency responses. Integrating multiple AI services efficiently within a single application. The importance of UX/UI design for accessibility in emergency scenarios.

What's next for BlazeGuard

Expanding support for more languages and accessibility features. Enhancing medical record security with encryption and user authentication. Integrating more advanced AI reasoning for complex medical inquiries. Exploring mobile app development for broader accessibility.