SECYRA AI

• 

  Secyra AI - App Icon

Inspiration

Digital threats are becoming part of everyday life. A suspicious SMS, a fake payment link, an unknown caller, a QR code, or a misleading website can look completely normal to an ordinary user. We wanted to build something that makes digital security easier to understand instead of expecting everyone to be a cybersecurity expert.

That idea inspired Secyra AI — a personal digital security companion designed to help users understand potentially risky digital situations and make safer decisions.

What it does

Secyra AI brings multiple security experiences into one Android application, including caller and number security, message analysis, link and QR security, device-security checks, privacy controls, security history, and AI-powered security assistance.

Its core approach is:

Observe → Analyze → Explain → Recommend → User decides

Rather than simply displaying a warning, Secyra is designed to explain what may be happening, why it could be risky, what evidence is available, how certain the result is, and what the user can do next.

AI-powered features are separated into the Pro experience, while the free version provides useful non-AI security functionality.

How I built it

Secyra AI is built as a native Android application using Kotlin and Jetpack Compose, with a Clean Architecture and MVVM structure.

The application is designed around a security-engine layer that can combine deterministic security checks, available threat intelligence, user context, and AI reasoning. AI requests are routed through a secure backend rather than exposing private API credentials inside the Android application.

We also designed the application around privacy and user control. Sensitive capabilities such as microphone access, camera access, and screen sharing are requested only when the corresponding feature needs them. Screen Security Session uses Android's explicit screen-capture consent mechanism.

RevenueCat is used for subscription and entitlement management, while Supabase is used for backend services and user data.

Challenges I faced

One of our biggest challenges was designing a security product without pretending that an AI model can automatically know whether everything is safe or dangerous. Security decisions require evidence, deterministic checks, uncertainty handling, and careful explanations.

Another challenge was working within Android's permission and privacy model. Features such as caller information, SMS access, microphone usage, camera access, and screen capture all have different technical and platform constraints.

We therefore designed Secyra around a simple principle: the user stays in control.

What I learned

Building Secyra taught us that good security software is not only about detecting threats. It is also about communicating risk clearly, protecting user privacy, handling uncertainty honestly, and making security understandable to people who are not cybersecurity professionals.

Our goal is to continue expanding Secyra into a broader personal security platform while keeping security, privacy, transparency, and user control at its core.

Built With

• ai
• android
• api
• compose
• cybersecurity
• jetpack
• kotlin
• learning
• llama
• machine
• privacy
• rest
• revenuecat
• sdk
• supabase