Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🛡️ Kavach AI — The Digital Armor for a Safer World

An AI-powered global guardian that protects people—especially elders—from online scams using empathy, language, cultural awareness, and intelligence.

Kavach AI Banner


📌 Project Overview

Kavach AI is a high-impact Streamlit web application built for global social good. It analyzes suspicious messages and screenshots to help users identify online scams—without using technical jargon—and adapts its communication based on the user’s country, language, and cultural context.

Instead of sounding like a cybersecurity tool, Kavach AI behaves like a trusted family member or personal guardian, guiding users calmly and clearly through risky digital situations.


🎯 Problem Statement

Online scams are a global problem affecting millions every day. People fall victim not because they lack intelligence, but because:

  • Scam messages create fear and urgency
  • Technical explanations are confusing
  • Help is not immediate, personal, or localized

Kavach AI solves this by combining AI intelligence with human empathy and cultural awareness.


✨ Key Features

🌍 Location-Aware Protection (Global)

  • Country / region selector
  • Scam understanding adapts to regional fraud patterns
  • Responses generated using native language, tone, and cultural norms
  • Provides local cybercrime reporting guidance based on location

🔍 Multimodal Scam Analysis

  • Text Analysis: Paste SMS, WhatsApp, Email, or social media messages
  • Image Analysis (OCR + Vision): Upload screenshots (JPG / PNG / JPEG)
  • Detects urgency, impersonation, fake branding, and malicious intent

🗣️ Native Language & Tone

  • Multilingual support with culturally appropriate phrasing
  • Avoids technical terms like phishing, malware, or HTTPS
  • Speaks gently, clearly, and reassuringly

“Please don’t click this link. It’s trying to scare you so you act fast. I’m here to keep you safe.”

🔊 Voice-First Friendly (Roadmap Ready)

  • Designed for text + voice output
  • Ideal for elders and non-technical users
  • Ready for future text-to-speech integration

🚦 Visual Risk Meter

  • 🟢 Safe
  • 🟡 Suspicious
  • 🔴 Dangerous

✅ Clear Actionable Guidance

Every analysis ends with exactly 3 simple steps, such as:

  1. Do not click or reply
  2. Block the sender
  3. Report to your local cybercrime authority

🖥️ Application Preview

Text Analysis Image Upload Risk Result
Text Input Image Upload Risk Meter

🧱 Tech Stack

Layer Technology
Frontend Streamlit (Python)
AI Engine Google Gemini 1.5 Flash
Multimodal Processing Gemini Vision + OCR
Localization Engine Country & Language Mapping
Styling Custom CSS
Deployment Streamlit Cloud / Cloud VM

🏗️ System Architecture

flowchart TB

    U[End User]

    U --> UI[Streamlit Frontend]

    UI --> TIN[Text Input]
    UI --> IIN[Image Upload]
    UI --> LOC[Country Selector]
    UI --> LANG[Language Selector]

    TIN --> IP[Input Processing]
    IIN --> IP

    IP --> TV[Text Validation]
    IP --> IV[Image Validation]

    IV --> ERR[Error Handling]

    TV --> CTX[Localization Engine]
    IV --> CTX
    LOC --> CTX
    LANG --> CTX

    CTX --> PROMPT[Prompt Engineering]

    PROMPT --> AI[Gemini 1.5 Flash API]

    AI --> RISK[Risk Classification]

    RISK --> SAFE[Safe]
    RISK --> SUSP[Suspicious]
    RISK --> DANG[Dangerous]

    SAFE --> RESP[Response Formatter]
    SUSP --> RESP
    DANG --> RESP

    RESP --> OUT[Localized Output]

    OUT --> UI


Loading

🔐 Security & Privacy

  • API keys are masked and stored only during runtime
  • No user messages or images are stored
  • All analysis is performed in real time

⚙️ Installation & Setup

# Clone the repository
git clone https://github.com/your-username/kavach-ai.git
cd kavach-ai

# Install dependencies
pip install -r requirements.txt

# Run the app
streamlit run app.py

🧪 Sample Use Cases

Fake bank or payment alerts

Urgent delivery or courier scams

Government or authority impersonation

Lottery, prize, or reward fraud

Social engineering messages on messaging apps


🌍 Global Social Impact

Protects users across countries, cultures, and languages

Makes cybersecurity understandable, not intimidating

Reduces fear-driven digital fraud

Encourages timely reporting and awareness

“Security should feel human—everywhere in the world. Because trust isn't built on code; it's built on understanding.”


🚀 Future Enhancements

Voice input & text-to-speech output

WhatsApp / Messenger bot integration

Country-specific scam trend dashboard

Family or caregiver alert system

Offline-first lightweight mode


⚙️ Scalability & Future-Ready Design

  • Plug-in support for:
    • Voice input / output
    • Messaging platform bots
    • Country-specific scam datasets
  • Can be deployed on:
    • Streamlit Cloud
    • Cloud VM
    • Containerized environments

🧠 Architectural Philosophy

Kavach AI is not just a detection system.
It is a communication system built on empathy, localization, and trust.


👨‍💻 Author & Vision

Kavach AI is built with the belief that technology should protect people, not overwhelm them.

If you find this project valuable, ⭐ the repository and help spread digital safety awareness.


📜 License

This project is licensed under the MIT License — free to use for innovation and social good.

About

No description, website, or topics provided.

Resources

Stars

2 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages