VidAlert

Inspiration

Imagine someone you love collapses right in front of you. Panic sets in. You call emergency services, but the ambulance is $8$ to $12$ minutes away. You know that every single minute without CPR strips away between $7\%$ and $10\%$ of their chance to survive. That "critical interval" is a desperate void where bystanders feel helpless, while certified responders—doctors or nurses living in your same building—have no idea someone is dying just yards away.VidAlert was born from that frustration. We refuse to accept a world where technology is used to deliver food in minutes but fails to save a life in seconds. We believe no one should die simply because the system failed to alert the person standing just around the corner.

What it does

VidAlert is an emergency response system that coordinates bystanders, certified responders, 144 operators, and ambulances in real-time. It allows witnesses to activate the system via WhatsApp or a mobile app. Once triggered, an adaptive Voice AI guides the witness through basic CPR step-by-step. Simultaneously, the system dispatches the most qualified nearby responder using an intelligent algorithm that considers proximity, medical specialty, and historical reliability.

How we built it

We developed a high-availability architecture using a modern tech stack:Frontend: Built with React Native and Expo for the mobile apps , and React + Vite for the 144 emergency dashboard.Backend: Powered by FastAPI (Python) to manage business logic and real-time validations.AI Engine: Utilized Claude API (Anthropic) to act as the "brain" for smart dispatching and the interactive CPR voice guide.Communications: Integrated Twilio for WhatsApp messaging and automated voice calls.Real-time Data: Used Firebase Firestore for live synchronization and Google Maps/Directions API for precise routing. Wearable & Health Integration: Developed natively in Android Studio to bypass limitations and directly monitor vital signs like heart rate and oxygen saturation.

Challenges we ran into

The biggest hurdle was the Intelligent Radius Expansion. We moved away from static search areas to a dynamic system. The challenge was orchestrating multiple variables in milliseconds—real-time traffic, the medic’s specific specialty, and their historical acceptance rate—to ensure the alert reaches the person who can actually make it to the scene in under $4$ minutes.

Accomplishments that we're proud of

Zero-Barrier Entry: Witnesses can activate the system via WhatsApp without needing to have any app installed.Privacy by Design: We implemented encrypted NFC tokens that only unlock a victim's medical history when a responder’s role is verified, an emergency is active, and their physical presence is confirmed via GPS.Explainable AI: Our 144 operator dashboard displays the specific "reasoning" behind every AI dispatch decision for full transparency.

What we learned

We learned to orchestrate complex cloud services for mission-critical scenarios where every second counts. We also gained deep insights into the importance of security audits and failsafe mechanisms ; if the AI API fails, the system automatically falls back to a distance-based logic to ensure the emergency is never ignored.

What's next for VidAlert

Our roadmap includes expanding the native Android implementation to support a wider range of hardware and developing predictive algorithms to identify high-risk zones, allowing us to position certified responders strategically before an emergency even occurs.