ScamSafe

• 

  Questionnaire

• 

  Home page

Inspiration

Online scams are becoming increasingly common and sophisticated. Many users either fail to recognize scams early or do not know what to do after something has already gone wrong. We wanted to build a tool that helps users both identify potential scams and respond effectively after a scam may have occurred, without overwhelming them or compromising privacy.

What it does

ScamSafe helps users in two ways:

1. Check if you’re being scammed
   Users answer a short questionnaire about a suspicious message, call, or email. The system provides a risk assessment and clear safety instructions. Users can optionally upload a screenshot for AI-assisted analysis.

2. I may have been scammed
   Users select what already happened (such as unauthorized money deduction or account compromise), and ScamSafe provides direct, action-oriented recovery steps.

The system combines rule-based logic with optional AI analysis to ensure reliability and explainability.

How we built it

• Frontend built using React, React Router, and Tailwind CSS
• Backend built using Node.js and Express
• Optional AI analysis using Google Gemini Vision
• Modular, question-driven architecture
• Privacy-first design with no data persistence

Challenges we ran into

• Designing flows for both scam detection and post-incident recovery
  
• Ensuring AI output was helpful without being misleading
  
• Managing multiple branching paths while keeping the UI simple
  
• Working under strict time constraints
  

Accomplishments that we're proud of

• Building a complete end-to-end system within a short timeframe
  
• Integrating AI responsibly without making it a single point of failure
  
• Providing clear, actionable guidance instead of generic warnings
  
• Designing separate preventive and recovery-focused flows
  
• Maintaining a strong focus on user privacy and consent
  

What we learned

• Rule-based systems and AI work best when clearly separated
  
• Simple UX is critical in high-stress situations
  
• AI output must be constrained to remain useful and trustworthy
  
• Clean architecture matters even under tight deadlines
  

What's next for ScamSafe

• Expanded scam scenarios and regional customization
  
• Multilingual support
  
• Mobile-first improvements
  
• Optional integrations for reporting scams
  

Built With

• api
• google-vision
• node.js
• react
• react-router
• tailwindcss