SCAMSHIELD AI

Cyber Defense Protocols — Link Neutralizer, Sender Blacklist and Sentiment Sensor modules
Threat Intelligence Summary — risk probability gauge with critical threat detection breakdown
Threat Intake Console — paste any SMS, email, chat or URL to run instant AI scam analysis
ScamShield AI landing page — real-time neural threat counter and quantum defense interface
AI processing screen — analyzing message vectors, urgency patterns and behavioral signals

Inspiration

We kept seeing it happen around us. A classmate's parent lost money to a fake bank SMS. A friend almost clicked a phishing link disguised as a college fee reminder. These weren't careless people — they were just caught off guard by messages designed to trick them.

We realized there was no simple tool where someone could just paste a suspicious message and get an instant answer. So we built one.

What it does

ScamShield AI lets anyone paste a suspicious SMS, email, chat message, or URL and get an instant AI-powered threat analysis.

The app returns a Risk Score from 0 to 100, identifies exactly what makes the message dangerous — malicious links, artificial urgency, credential harvesting patterns, impersonation tactics — and displays everything in a clear threat dashboard.

No technical knowledge needed. Just paste and scan.

How we built it

We split the work across three modules.

Tejaswi handled the entire frontend — the futuristic interface, particle animations, scanning effects, custom cursor, and the hero section.

Swathi built the AI analysis engine — keyword detection, behavioral pattern scoring, risk score calculation, and threat vector analysis.

Rishika built the results dashboard, threat logs, defense protocols section, and the PDF export system.

The whole app runs on HTML, CSS, and Vanilla JavaScript — no backend, no database, fully client-side.

Challenges we ran into

Getting the AI scoring to feel genuinely intelligent without a real ML backend was the hardest part. We had to design a keyword and pattern detection system that actually caught real scam tactics — urgency language, suspicious domains, impersonation phrases — not just random words.

Typography was also tricky. The futuristic fonts and layout kept breaking at different screen sizes and we spent a lot of time fixing that.

What we learned

How real scam messages are structured and what psychological tactics they use. How to build a convincing AI analysis pipeline on the frontend. And honestly — how to work as a team under pressure and ship something we're actually proud of.