Shecurity

Shecurity is an AI-powered safety app disguised as a calculator, offering scream detection, discreet SOS alerts, live location sharing, and secure audio evidence preservation.

Comment

Inspiration

Women often find themselves in situations where openly accessing a safety application or pressing an SOS button is difficult or even dangerous. We wanted to create a solution that works discreetly, responds to potential threats, and helps preserve crucial evidence. This inspired us to build Shecurity — a safety companion designed to make protection more accessible when every second matters.

What it does

Shecurity is an AI-powered personal safety application disguised as a regular calculator. Its key features include:

Calculator Disguise: Provides a familiar calculator interface with a hidden safety dashboard.

AI Scream Detection: Uses sound classification to identify potential distress signals.

Discreet SOS: Allows users to trigger emergency assistance without drawing attention.

Duress PIN: Creates a fake cancellation experience while silently escalating an emergency.

Live Location Sharing: Helps users share their location with trusted contacts.

Audio Black Box: Captures audio before and after an emergency trigger to help preserve important evidence.

Evidence Vault: Stores recordings locally, generates SHA-256 hashes, and creates downloadable PDF reports.

How we built it

We developed Shecurity using React, TypeScript, and Vite, with Tailwind CSS for styling and Framer Motion for smooth animations.

For AI-powered sound classification, we integrated TensorFlow.js with Google's YAMNet model. The Web Audio API and AudioWorklet handle microphone input and audio processing. Browser Geolocation provides location tracking, while Leaflet and OpenStreetMap support map functionality.

We used IndexedDB for local evidence storage, Web Crypto API for cryptographic hashing and password protection, and jsPDF for generating evidence reports. The application is deployed on Netlify.

Challenges we ran into

Integrating real-time AI audio classification without affecting application performance.

Distinguishing potential distress sounds from everyday background noise.

Managing continuous audio buffering while preserving relevant emergency recordings.

Implementing discreet emergency activation and duress PIN functionality.

Handling browser restrictions involving microphone permissions, background execution, GPS access, and emergency communication.

Balancing user privacy, evidence security, and ease of use.

Accomplishments that we're proud of

Building a functional safety application with a disguised calculator interface.

Integrating an AI-based sound classification pipeline.

Developing a discreet emergency escalation mechanism with a duress PIN.

Implementing an audio evidence vault with SHA-256 integrity checks.

Combining location sharing, emergency alerts, and evidence preservation in one application.

Creating a complete, responsive web application with a distinctive user experience.

What we learned

This project helped us understand the practical challenges of building AI-powered applications, real-time audio processing, browser APIs, and privacy-conscious systems. We learned that developing a safety application requires more than implementing features — reliability, accessibility, user consent, and responsible handling of sensitive information are equally important.

What's next for Shecurity

We aim to evolve Shecurity into a more reliable, accessible, and comprehensive personal safety platform by:

Developing a native Android application with background monitoring capabilities.

Improving AI accuracy and reducing false-positive detections.

Introducing multilingual distress detection.

Integrating automated emergency notification services through authorized channels.

Strengthening evidence security with trusted timestamps and digital signatures.

Adding hardware-based emergency triggers and wearable device integration.

Expanding accessibility and personalization for different user groups.

Built With

Share this project:

Updates

Submission history