SALVUS

• 

  https://drive.google.com/file/d/1oImcOk_AwX0p1pa8oBZa3Az4KB4kxSJ8/view?usp=sharing

Salvus was inspired by the need for faster, smarter disaster response in regions where minutes can save lives. We wanted to move from reactive alerts to predictive intelligence using AI. We built a real-time system that collects environmental data, analyzes risk, verifies emergency reports, and guides responders with clear recommendations. Along the way, we learned the importance of explainable AI, real-time system design, and offline resilience. Our biggest challenges were modeling compound risks, preventing alert fatigue, and keeping the system reliable under dynamic conditions.

Built With

• css
• elevenlabs-api
• fastapi
• gemini-api
• html
• javascript
• leaflet.js
• open-meteo-api
• openaq-api
• openstreetmap
• progressive-web-app-(pwa)
• python
• react
• recharts
• server-sent-events-(sse)
• sqlite
• tailwind-css
• twilio
• usgs-earthquake-api
• vercel