Scam Sense

Learn, practice, and master scam detection skills

Comment

Inspiration

In today's digital age, scams and fraudulent content are becoming increasingly sophisticated, targeting vulnerable populations who may lack the technical literacy to identify threats. We were inspired to create TechLit after witnessing how easily people fall victim to phishing emails, fake news, and investment scams. Our goal was to bridge the digital literacy gap by providing an accessible, interactive platform that teaches users to recognize and avoid fraudulent content.

What it does

ScamSense is an interactive fraud detection education platform that helps users learn to identify scams and fraudulent content.

Key Features:

AI-Powered Analysis - Users can paste suspicious emails, texts, or messages and get instant fraud risk assessment using AWS Bedrock AI

Daily Challenges - Gamified learning with daily multiple-choice questions about real scam scenarios

Streak Tracking - Users build streaks by completing daily challenges, encouraging consistent learning

Educational Explanations - Detailed breakdowns of why content is fraudulent, teaching users to spot red flags

User Accounts - Personal progress tracking and streak history

How it works: Users register, analyze suspicious content for immediate feedback, complete daily challenges to build knowledge, and track their learning progress through streaks. The platform combines AI intelligence with educational content to make cybersecurity awareness accessible and engaging for everyone.

How we built it

Technology Stack: Frontend: React.js with modern CSS for responsive design

Backend: Flask with CORS support for API endpoints

Database: SQLite with SQLAlchemy ORM

AI: AWS Bedrock with Meta Llama 3 8B Instruct model

Authentication: JWT tokens for secure user sessions

Challenges we ran into

  1. AWS Bedrock Integration Challenge: Initial setup and authentication with AWS services Solution: Implemented proper credential management and fallback mechanisms

  2. Database Schema Evolution Challenge: Adding streak tracking required database migrations Solution: Used db.drop_all() and db.create_all() for development, learned about proper migration strategies

  3. Frontend-Backend Communication Challenge: CORS issues and proxy configuration between React and Flask Solution: Configured proper CORS headers and used full URLs when proxy failed

  4. State Management Challenge: Keeping navbar streak updated after daily challenge completion Solution: Implemented custom events and localStorage synchronization:

  5. User Authentication Flow Challenge: JWT token validation was causing 422 errors Solution: Simplified authentication for development while maintaining security principles

  6. Daily Challenge Logic Challenge: Ensuring each user gets their own streak tracking Solution: Implemented proper user identification and database relationships

Accomplishments that we're proud of

Successfully integrated AWS Bedrock AI with Meta Llama 3 for intelligent fraud detection

Built a complete full-stack application with React, Flask, and SQLite from scratch

Created a gamified daily challenge system with streak tracking to encourage learning

Implemented secure user authentication with JWT tokens and password hashing

Designed responsive, accessible UI that works seamlessly across all devices

Developed smart pattern recognition for detecting phishing, scams, and fraudulent content

Achieved real-time analysis with instant feedback for users

Built robust error handling with AI fallback mechanisms for reliability

What we learned

Throughout this project, we gained valuable experience in:

Full-stack development with React frontend and Flask backend

AI integration using AWS Bedrock and Meta Llama models for intelligent fraud analysis

User authentication and session management with JWT tokens

Database design with SQLAlchemy for user data and streak tracking

Real-time analysis of text patterns using both AI and rule-based approaches

Gamification principles through daily challenges and streak systems

What's next for Scam Sense

More Advanced AI Models: Integration with more sophisticated fraud detection models

Social Features: Leaderboards and community challenges

Content Expansion: More diverse scam types and scenarios

Analytics Dashboard: User progress tracking and insights

Built With

Share this project:

Updates