🧠 NeuroNews — AI-Enhanced News Platform

NeuroNews is a full-stack, AI-powered news platform built as my Semester 6 BSc Computer Science project.

The goal was to create a smarter way to discover, track, and engage with news using modern backend architecture and AI-driven features.

This project was designed and developed independently, covering everything from backend APIs and authentication to real-time systems and AI model integration.

⚠️ Note: The source code for this project is private. This repository is intended to document the design, features, and technical learnings.

---

🎯 Project Objectives

• Build a scalable MERN-based news platform with secure authentication
• Design robust REST APIs and real-time communication
• Integrate AI models for intelligent content processing
• Gain end-to-end full-stack development experience

---

🚀 Key Features

🔐 Secure Authentication

• OTP-based user registration and login
• Email verification using Nodemailer
• JWT-based authentication for login and all client requests

---

🏠 Intelligent News Dashboard

• Homepage displaying top headlines
• Category-based filtering (Science, Sports, Technology, etc.)
• Data aggregated from multiple third-party News APIs

---

🔍 Smart Search & News Reminders

• Search news using natural language prompts

Example:
“Tell me if there’s news about Tata stock”

• Users receive email notifications when relevant news appears later
• Dedicated reminders section to view and manage saved alerts

---

📬 Reminder Management

• View all saved reminders
• Delete reminders directly from the dashboard
• Structured storage for fast access

---

🌐 Chrome Extension + Real-Time System

When a user opens a news article on the original publisher’s website:

• A custom Chrome extension activates
• Establishes a WebSocket connection with the backend
• Backend uses web scraping to fetch related YouTube videos in real time
• Enhances user engagement without leaving the news context

---

🤖 AI-Powered News Categorization

• Many news APIs lack proper category tagging
• Implemented Facebook BART model via Flask (Python) to classify news articles based on content
• Enabled accurate categorization across all sources

---

🧠 Keyword Extraction & Optimized Storage

• Extracted relevant keywords from articles
• Stored news in MongoDB, organized by:
  • Date
  • Category
  • Keywords
• Designed for fast querying and scalability

---

🛠️ Tech Stack

Frontend

• React.js

Backend

• Node.js
• Express.js
• MongoDB

AI & Data Processing

• Flask (Python)
• Facebook BART Model

Real-Time & Automation

• WebSockets
• Puppeteer (Web Scraping)

Other Tools

• Nodemailer
• JWT Authentication
• Chrome Extension

---

3 Third-Party News APIs Used

• News.org
• NewsAPI
• GNews API

---

🧩 System Architecture (High-Level)

• MERN stack for core application
• Flask Server for AI-based categorization
• WebSockets for real-time browser–backend communication
• MongoDB for structured, efficient data storage

---

📸 Project Screenshots

Homepage

---

Cetegories Menu

---

Specific Category Page (Science)

---

Chrome extension in action

---

📚 Key Learnings

• Backend architecture and REST API design
• Secure authentication and authorization
• Real-time communication using WebSockets
• Integrating AI models into full-stack applications
• Designing systems for scalability and maintainability
• Working independently on a production-like project