news-digest

• 

• 

• 

• 

• 

• 

Inspiration

The idea came from a very real problem I faced daily — information overload. There’s too much news online, and most of it is either repetitive, irrelevant, or time-consuming to read.

I wanted to build something that:

filters out noise delivers only what matters and does it in the most efficient way possible

That’s where the idea of a personalized AI-powered news digest came from — something that works quietly in the background and gives you exactly what you need.

What it does

This system acts like a personal news assistant.

Users subscribe with their email Choose topics they care about (tech, AI, etc.) Select frequency (daily/weekly)

Behind the scenes, the system:

Scrapes news from multiple sources Uses AI (Gemini) to summarize articles Performs sentiment analysis (positive / negative / neutral) Stores everything in a database Sends a clean, personalized email digest automatically

Users can:

Browse summarized articles on a dashboard Update preferences anytime Unsubscribe securely via tokenized links

How we built it

I built this as a full-stack system, focusing on modularity and scalability.

Backend:

Node.js + Express for APIs MongoDB for storing users and articles Axios + Cheerio for scraping Gemini API for summarization Nodemailer for email delivery node-cron for scheduling automation

Frontend:

React (Vite) for fast UI development Tailwind CSS for clean design Axios for API calls React Router for navigation

Architecture flow: Scraping → AI Processing → Database → Email Delivery → Frontend Display

Everything is decoupled so each part can scale independently.

Challenges we ran into

One of the biggest challenges was handling the pipeline flow:

Scraping → summarizing → sentiment → storing → emailing Each step depends on the previous one, so handling failures and async execution properly was tricky.

Other challenges:

Avoiding duplicate or low-quality scraped content Managing API rate limits (AI + scraping sources) Designing a clean unsubscribe system with secure tokens Keeping email formatting consistent across devices

Accomplishments that we're proud of

Built a complete end-to-end AI product, not just a feature Successfully automated the entire pipeline (no manual work needed) Integrated AI in a practical, real-world use case Designed a clean UI that actually makes news consumption easier Created a system that can realistically scale into a SaaS product

What we learned

How to integrate AI meaningfully instead of just adding it as a feature Managing asynchronous workflows in real systems Designing APIs that are clean and scalable Structuring a full-stack project properly Thinking like a product builder, not just a developer

What's next for news-digest

This project has strong potential to evolve further. Planned improvements:

Add authentication (JWT-based user accounts) Smarter personalization using user behavior (not just selected topics) Push notifications (mobile + web) Search and bookmarking features AI-powered recommendations (“you might also like”) Multi-language support Convert into a full SaaS platform with analytics dashboard

Built With

• axios
• cheerio
• express.js
• google-gemini-api
• javascript-(es6+)
• mongodb
• node-cron
• node.js
• nodemailer
• react-(vite)
• react-router
• rest
• tailwind-css
• vercel-(frontend-deployment)