-
-
Types of Scams
-
Paste Message Section
-
Add API key option
-
Landing Hero Section
-
Report
📌 About the Project
💡 Inspiration
Scams like phishing, fake alerts, and fraud messages are increasing rapidly. Most users cannot quickly verify authenticity.
This project was built to create a simple, fast, and intelligent scam detection tool.
🛠️ How We Built It
We used:
- Streamlit for UI
- AI (LLM) for understanding message intent
- Rule-based checks for known scam patterns
Risk score calculation:
$$ Risk\ Score = \min(100,\ \sum w_i \cdot f_i(x)) $$
Where ( f_i(x) ) represents scam signals like urgency, fake links, etc.
AI-based decision:
\( Verdict = f_{AI}(context,\ intent,\ patterns) \)
🧠 What We Learned
- Prompt engineering for structured AI output
- Importance of explainable AI
- Building fast MVP under time constraints
⚔️ Challenges
- Limited time (~3 hours)
- Maintaining consistent AI output
- Handling multiple types of scams
🌍 Impact
The system reduces scam success probability:
$$ P(Success) \propto \frac{1}{Awareness} $$
As awareness increases:
$$ P(Success) \downarrow $$
🚀 Conclusion
A simple AI tool that helps users instantly detect scams and stay safe online.
Built With
- javascript
- next.js
- node.js
- openai
- react
