🛡️ TRINETRA

AI-Powered Digital Public Safety Intelligence Platform

Detect. Verify. Protect.

---

🇮🇳 Protecting India from Digital Fraud using AI

Trinetra is an AI-powered Digital Public Safety Intelligence Platform built for ET AI Hackathon 2.0 (Problem Statement 6).

Instead of detecting fraud after citizens lose money, Trinetra helps detect scams before financial loss occurs using Artificial Intelligence.

It provides real-time analysis for:

• 🚨 Digital Arrest Scams
• 💬 Fraudulent WhatsApp/SMS Messages
• 📄 Fake Government & Employment Documents
• 🌍 Community Scam Intelligence

---

🎯 Problem Statement

India is witnessing an alarming rise in

• Digital Arrest Scams
• Fake Job Scams
• Banking & KYC Fraud
• Fake Government Notices
• AI-powered Social Engineering

Traditional systems are reactive.

Trinetra shifts fraud detection from reactive investigation to proactive prevention.

---

✨ Features

🚨 Digital Arrest Scam Detector

Analyze suspicious call transcripts, SMS or WhatsApp chats.

✔ CBI / ED / Police impersonation detection

✔ Psychological pressure detection

✔ Financial demand detection

✔ Isolation tactic detection

✔ Multilingual support

---

💬 Message Fraud Detector

Detects

• Fake Jobs
• Investment Scams
• Lottery Scams
• Task Scams
• KYC Fraud
• Romance Fraud
• Government Impersonation

Returns

• Risk Score
• Scam Category
• Red Flags
• Citizen Guidance

---

📄 AI Document Verifier

Upload PDF documents.

Trinetra checks for

• Fake Government Notices
• Fake Arrest Warrants
• Fake Offer Letters
• Fake Banking Documents
• Fake Customs Notices

---

🌍 Community Scam Intelligence

Citizens can report scams.

The platform builds a live community intelligence feed powered by Supabase.

---

📑 Audit Trail

Every AI analysis generates

• Unique Case ID
• UTC Timestamp
• Platform Signature

making reports easier to reference.

---

🧠 AI Pipeline

Citizen Input
      │
      ▼
React Frontend
      │
      ▼
FastAPI Backend
      │
      ▼
Llama 3.3 70B (Groq)
      │
      ▼
Risk Analysis
      │
      ▼
Deterministic Verdict Engine
      │
      ▼
Citizen Guidance

---

⚙️ Tech Stack

Layer	Technology
Frontend	React 18 + Tailwind CSS + Vite
Backend	FastAPI + Python
AI	Llama 3.3 70B via Groq
Database	Supabase PostgreSQL
Deployment	Netlify + Render
APIs	REST

---

🎯 Risk Scoring

Digital Arrest

Score	Verdict
0–20	🟢 SAFE
21–45	🟡 LOW RISK
46–70	🟠 MODERATE
71–89	🔴 HIGH RISK
90–100	🚨 CONFIRMED SCAM

---

Message Fraud

Score	Verdict
0–20	SAFE
21–65	Suspicious
66–85	Likely Scam
86–100	Definite Scam

---

Document Verification

Score	Verdict
0–25	Legitimate
26–50	Suspicious
51–75	Likely Fraud
76–100	Confirmed Fraud

---

📊 Architecture

                React Frontend
                       │
             REST API Requests
                       │
                 FastAPI Backend
        ┌──────────────┼──────────────┐
        │              │              │
        ▼              ▼              ▼
Digital Arrest   Message Fraud   Document Verify
        │              │              │
        └──────────────┼──────────────┘
                       ▼
             Llama 3.3 70B (Groq)
                       │
             Deterministic Verdict
                       │
             Audit Trail Generation
                       │
                  Supabase Database

---

🚀 Local Setup

git clone https://github.com/RithishPK/trinetra.git

cd trinetra

Backend

cd backend

python -m venv venv

pip install -r requirements.txt

uvicorn main:app --reload

---

Frontend

cd frontend

npm install

npm run dev

---

🔑 Environment Variables

GROQ_API_KEY=

SUPABASE_URL=

SUPABASE_KEY=

---

📷 Screenshots

Add screenshots here

Landing Page

Digital Arrest Detector

Message Fraud Detector

Document Verifier

Community Feed

---

🏆 Built For

ET AI Hackathon 2.0

Problem Statement 6

AI for Digital Public Safety

---

👨‍💻 Team

Rithish PK

Chinmay Deepak Chandavar

---

⭐ If you found this project useful, consider giving it a star!

Made with ❤️ for a safer digital India 🇮🇳