GuardianAI

GuardianAI is an AI-powered scam detection system that analyzes emails, text, and images to detect phishing and online scams using Google Gemini AI.

Comment

Inspiration

Online scams and phishing attacks are increasing every day. Many users cannot easily identify fraudulent emails and messages. We wanted to build an AI-powered assistant that helps users detect scams quickly and safely.

What it does

GuardianAI analyzes suspicious text, emails and screenshots using Google's Gemini AI. It provides:

  • Scam risk score
  • Threat level
  • Scam category
  • Confidence score
  • Reasons behind the prediction
  • Safety recommendation

The system also extracts text from uploaded images using OCR before performing AI analysis.

How we built it

Backend:

  • Python
  • FastAPI
  • Google Gemini AI API
  • EasyOCR

Frontend:

  • React
  • JavaScript
  • HTML
  • CSS

The frontend communicates with the FastAPI backend through REST APIs.

Challenges we ran into

  • Setting up the Gemini API correctly
  • Handling OCR from uploaded images
  • Connecting the React frontend with FastAPI
  • Managing API responses and displaying structured results

Accomplishments that we're proud of

  • Successfully integrated Gemini AI
  • Built a complete frontend and backend
  • Added image-based scam detection
  • Generated detailed AI explanations instead of only Yes/No predictions

What we learned

  • FastAPI development
  • React integration
  • Gemini AI API
  • OCR processing
  • Building AI-powered full-stack applications

What's next for GuardianAI

  • Browser extension
  • Mobile application
  • QR code scam detection
  • Email inbox integration
  • Multi-language support

Built With

Share this project:

Updates

Submission history