-
Don't get scammed — get ScamSentry. Your digital defense starts here.
🎯 Inspiration
Scams are everywhere — phishing links, fake bank calls, WhatsApp investment traps — and they’re getting smarter every day. What shocked us was how easily people fall for them, especially elders, students, and rural users who may not be digitally literate.
A close relative losing money to a scam became the turning point for us. We realized there’s a critical gap in scam awareness, detection, and reporting, and technology could fill it.
We wanted to build something that goes beyond static awareness — something smart, real-time, and accessible to anyone in any language. That’s how ScamSage was born.
💡 What it does
ScamSentry is an AI-powered, community-driven platform to:
- 🧠 Analyze suspicious messages with ScamGPT (AI-based scam analyzer)
- 🔎 Report scams anonymously and securely
- 🌐 Visualize scam hotspots with live maps and trend dashboards
- 🎙️ Educate users with multilingual articles, voice narration, and daily scam alerts
- 🧱 Log scams on blockchain for tamper-proof, legal-grade evidence
Our goal: Prevent scams before they strike and create a safer digital world.
🛠️ How we built it
We built the platform using modern, scalable technologies:
| Component | Technology |
|---|---|
| Frontend | React.js + Tailwind CSS |
| Backend | Python (FastAPI) |
| AI Integration | OpenAI GPT-4 API + Custom heuristics |
| Database | PostgreSQL |
| Maps | Leaflet.js + OpenStreetMap |
| Blockchain | IPFS + Polygon (via Web3.py) |
| Multilingual | Google Translate API + gTTS |
We started with the AI analyzer, then built the reporting form, mapped live data, and finally added blockchain logging and regional language support.
🚧 Challenges we ran into
- ⚠️ Data Scarcity: Real scam messages are often confidential or deleted. We had to simulate and categorize thousands of samples manually.
- 🧠 AI Prompt Tuning: Making GPT give consistent and accurate responses about scam likelihood was tricky.
- 🌐 Regional Language Integration: Translating and voicing educational content for multiple languages required proper encoding and storage.
- ⛓️ Blockchain Optimization: Logging to IPFS/Polygon cost gas fees — we optimized the payload and batching.
🏆 Accomplishments that we're proud of
- ✅ Built a working AI-based scam detection tool
- ✅ Created a real-time scam heatmap dashboard
- ✅ Integrated blockchain logging of scam reports
- ✅ Made a multilingual, voice-accessible awareness portal
- ✅ Ran test cases with fake scam messages and got 85%+ accuracy
But most importantly, we made something that could truly help people, even beyond the hackathon.
📚 What we learned
- How to combine AI + Web3 + Maps into one impactful solution
- Real-world UX matters: simple layouts and audio can change accessibility
- Handling language diversity in India is tough but doable
- People are more willing to report if they feel safe and anonymous
We also learned the importance of data ethics — protecting user privacy while collecting and using sensitive information.
🚀 What's next for Scamsentry – The Scam Intelligence & Prevention Network
We plan to:
- 📱 Launch a mobile app for rural access
- 🤖 Build a WhatsApp & Telegram bot version of Scam Analyzer
- 🌍 Partner with local cybercrime units for verified reports
- 🔐 Add scam report verification via SMS OTP or Aadhaar (optional)
- 🧑🏫 Conduct community workshops + run awareness drives in regional schools and banks
Ultimately, we envision ScamSentry as the digital firewall for the masses — empowering users to fight scams before they spread.
Built With
- and-integrated-openai's-api-for-analyzing-suspicious-text-and-detecting-scam-patterns.-the-ui-was-styled-using-bootstrap
- and-javascript-for-the-frontend
- and-python-with-flask-for-the-backend.-we-used-sqlite-as-the-database-for-storing-scam-reports
- css
- scamsage-was-built-using-html
Log in or sign up for Devpost to join the conversation.