KwaniGuard

KwaniGuard uses AI to detect scam messages in real-time. Inspired by Swahili 'kwani' (why?), it questions suspicious texts/images like a digital bodyguard.

Comment

Inspiration

With scam messages becoming increasingly sophisticated, we built ScamShield to help people instantly verify suspicious content.

What It Does

Analyzes text messages or uploaded images using AI to detect:

  • Phishing attempts
  • Financial scams
  • Fake prize offers
  • Suspicious links

How We Built It

  • Frontend: HTML/CSS/JS
  • Backend: Python/Flask
  • APIs: Google Cloud Vision + Gemini AI
  • Deployment: Render (free tier)

Challenges

  • Tuning AI prompts for accurate scam detection
  • Handling image-to-text conversion reliably

Accomplishments

  • Created a working prototype in 48 hours
  • Achieved 85%+ accuracy on test scam samples

What's Next

  • Add multilingual support
  • Browser extension version
  • Community scam database

Built With

Share this project:

Updates