Kavach AI

Moving from reactive safety to predictive protection.

Comment

Inspiration

It all started when I was heading out at 5:00 AM to give my JEE exam. Seeing me outside so early, a few people stopped to ask about my safety and why I was alone at that hour. That exact moment sparked a deep realization: public spaces are still filled with anxiety for women, even when they are just chasing their dreams. Existing apps only help after an emergency occurs. I wanted to build something that prevents the danger entirely, ensuring no girl feels unsafe, no matter the time.

What it does

Kavach AI is a proactive, predictive women's safety system that shifts security from reactive to preventive.It starts with secure Biometric & Voice Setup to establish user identity.Using real-time data (location, time, and movement patterns), its AI Risk Decision Engine calculates a dynamic risk percentage and suggests the safest routes.If a user deviates from the route or fails to share their destination, the system triggers a Periodic Safety Check-In.In high-risk scenarios, or if the phone is forcefully powered off, an Intelligent Silent Response System automatically broadcasts the live location and SOS alerts to emergency contacts and nearby public alarm systems.

How we built it

Architecture & Workflow: Designed the complete systematic backend logic and user journey flow using digital prototyping tools.Risk Calculation Logic: Structured a theoretical AI predictive model that evaluates danger levels (Low, Medium, High) based on spatial and temporal factors.Proposed Integrations: Mapped out the integration of Google Maps API for route deviation tracking and Twilio API for automated silent SOS broadcasting.

Challenges we ran into

The Power-Off Dilemma: One major challenge was figuring out how the system would protect a user if an attacker forcefully turned off their phone. We solved this by designing a server-side backend trigger that activates an automatic SOS if the device suddenly disconnects without a user's secure PIN authentication.Preventing False Alarms: Differentiating between a genuine threat and a simple accidental route change was tough. We handled this by adding a quick periodic safety check-in prompt before escalating the risk level.

Accomplishments that we're proud of

Creating a unique Predictive Safety Architecture that focuses on danger prevention rather than just post-incident alerts.Designing a robust, foolproof workflow that works even under extreme conditions like network loss or device shutdown.Translating a real-life personal experience from a 5:00 AM exam morning into a fully thought-out tech solution to protect millions of women.

What we learned

We learned how to deeply analyze a real-world social problem and break it down into systematic, logical technical blocks.We gained a solid understanding of how AI engines can use everyday data points (like time and location coordinates) to predict human safety metrics.

What's next for Kavach AI

Working Prototype: Developing the actual mobile application using Flutter and integrating live machine learning models.Offline Mesh Network: Adding Bluetooth and SMS gateway mesh networks so the app can send distress signals even in zero-internet zones.Community Integration: Partnering with local authorities, public transport networks, and smart city alarm infrastructure to make the response time lightning fast.

Hybrid capabilities and offline fallback

To ensure 100% reliability, Kavach AI functions as a complete hybrid system designed to operate under all network conditions:Online System: When internet connectivity is available, the app utilizes its core AI Predictive Engine to monitor real-time route deviations and calculate live safety risk scores.Offline System (No Internet Mode): In remote or rural areas with zero internet coverage, the system automatically switches to an offline backup mode. It leverages Bluetooth Mesh Communication to establish short-range alerts with nearby devices.Emergency SMS Fallback: If a high-risk anomaly or panic state is detected without data connectivity, the system bypasses the internet to send automated emergency SOS alerts via encrypted offline SMS gateways.This ensures that Kavach AI remains a smart, secure, and practical solution for low-income users and rural environments where emergency safety is needed the most.

Built With

Share this project:

Updates