SafeSphere AI 🛡️

AI-Powered Cyberbullying Detection & Reporting Platform

About

SafeSphere AI is a full-stack web platform that detects harmful content in real time and provides a structured complaint filing and resolution system. Built to give cyberbullying victims a voice and ensure 24-hour admin response.

Features

• Real-Time AI Detection — 400+ keyword detection with severity scoring (low/medium/high/critical)
• 3-Step Complaint Filing — Guided form with AI severity detection and image evidence upload
• Admin Control Center — Review, assign, resolve complaints with email notifications
• User Dashboard — Safety score, analysis history, complaint tracking
• Google Sign-In — OAuth 2.0 integration
• OTP Password Reset — Secure email OTP via Resend API
• Role-Based Access — User, Admin roles with protected routes

Tech Stack

Layer	Technology
Frontend	React.js, Tailwind CSS, Framer Motion
Backend	Node.js, Express.js, REST API
Database	MongoDB Atlas
Auth	JWT, Google OAuth 2.0
Email	Resend API
Deployment	Vercel + Render

Getting Started

Prerequisites

• Node.js 18+
• MongoDB Atlas account
• Resend API key

Installation

# Clone the repo
git clone https://github.com/deekshitha875/safesphere-ai.git
cd safesphere-ai

# Install server dependencies
cd server && npm install

# Install client dependencies
cd ../client && npm install

Environment Variables

Create server/.env:

PORT=5001
MONGO_URI=your_mongodb_uri
JWT_SECRET=your_jwt_secret
RESEND_API_KEY=your_resend_key

Create client/.env:

VITE_API_URL=http://localhost:5001
VITE_GOOGLE_CLIENT_ID=your_google_client_id

Run Locally

# Terminal 1 - Backend
cd server && npm start

# Terminal 2 - Frontend
cd client && npm run dev

Live Demo

Visit safesphere-ai-gamma.vercel.app

License

MIT