What Is This?
An intelligent news aggregation system that automatically collects articles from multiple sources, filters them based on your interests, summarizes them in your preferred language, and consolidates duplicate stories—all while you sleep.
Unlike traditional RSS readers or news aggregators, this system uses AI to understand your interests, not just match keywords. It reads full articles, identifies what matters to you, and presents a clean, consolidated view of your personalized news landscape.
🎯 Smart Interest Filtering
Uses GPT-4o-mini to understand article content and match it against your nuanced interests. Not just keyword matching—actual comprehension.
🌍 Multi-Language Support
Reads articles in any language, translates and summarizes them in your preferred language. Perfect for expats and international news followers.
🔄 Automatic Consolidation
When multiple sources cover the same story, AI merges them into a single comprehensive summary. No more reading the same news twice.
📱 Mobile-First Design
Responsive, dark-mode interface optimized for reading on any device. Print-friendly light mode automatically activates when needed.
🔒 Privacy Focused
Runs on your own server. Your reading habits and interests stay private. No tracking, no ads, no data selling.
⚡ Zero Maintenance
Set it up once, and it runs automatically every few hours. Git-based deployment with automatic updates to your hosting platform.
How It Works
1. Configure Your Sources
Add any news website to your websites.json file. Specify CSS selectors for article cards, titles, and summaries. Support for sites like Hacker News, news portals, blogs, and more.
2. Define Your Interests
For each source, describe what topics matter to you in natural language. The AI understands context and nuance—no need for rigid keyword lists.
3. Let AI Do the Work
Every few hours, the system:
- Fetches articles from all your sources
- Filters them against your interests using AI
- Reads and summarizes matching articles
- Consolidates duplicate stories
- Generates a clean HTML page
- Commits and deploys automatically
4. Read Your Digest
Access your personalized news digest anytime, from anywhere. Each day gets its own page with an auto-generated index for easy navigation.
Built With Modern Tech
Perfect For
🌐 Expats & International Citizens
Stay informed about your home country while living abroad. Track news that affects dual citizens, visa requirements, and international policies—all translated to your preferred language.
💼 Engineering Leaders
Keep up with software engineering trends, developer tools, programming languages, and tech policy without drowning in Hacker News. AI filters the signal from the noise.
📰 News Junkies
Follow multiple news sources without reading the same story ten times. AI consolidates coverage from different angles into coherent summaries.
🔬 Researchers & Specialists
Monitor niche topics across multiple publications. Medical breakthroughs, physics discoveries, space news—whatever your field, stay current effortlessly.
Key Features
- Intelligent Filtering: AI reads articles and understands context, not just keywords
- Cross-Language Support: Read sources in any language, get summaries in yours
- Duplicate Detection: Smart consolidation merges similar stories from multiple sources
- SQLite Tracking: Never process the same article twice
- Automatic Deployment: Git-based workflow auto-deploys to Netlify or similar platforms
- Responsive Design: Beautiful reading experience on desktop, tablet, and mobile
- Dark Mode Default: Easy on the eyes with automatic light mode for printing
- Error Resilient: Continues processing even if individual articles fail
- Comprehensive Logging: Full audit trail of all processing activities
- No Dependencies on APIs: Uses curl for fetching, works with any website
Example Use Cases
🇺🇦 Ukrainian Expat in Netherlands
Sources: Dutch immigration news, Ukrainian war updates, Kryvyi Rih regional news
Interests: Conscription laws, dual citizenship, attacks on home region, visa requirements
Result: Daily digest in Dutch covering everything that matters without overwhelm
💻 Software Engineering Manager
Sources: Hacker News, Popular Mechanics, tech blogs
Interests: Developer productivity, macOS tools, AI coding assistants, team building
Result: Curated feed of relevant tech news without the noise
🔬 Medical Professional
Sources: Medical journals, science news sites
Interests: Breakthrough treatments, clinical trials, medical technology
Result: Concise summaries of latest medical advances in your specialty
Why This Approach?
Traditional news aggregators fail because they rely on keywords and RSS feeds. They can't understand nuance, can't read paywalled content, and can't consolidate stories from multiple sources.
This system uses AI for every step:
- Fetches full article content (not just headlines)
- Understands your interests in context
- Reads and comprehends article meaning
- Translates and summarizes intelligently
- Identifies duplicate coverage
- Merges related stories into comprehensive summaries
The result? Signal, not noise.
Open Source & Self-Hosted
This system is designed to be:
- Self-hosted: Run on your own server or serverless platform
- Customizable: Full control over sources, interests, and output format
- Private: Your data stays yours
- Extensible: Built with clean, modular code that's easy to modify
- Cost-effective: Uses GPT-4o-mini for optimal balance of quality and cost
Get Started
Setting up your own instance takes less than 30 minutes:
- Clone the repository
- Install dependencies (
npm install) - Add your OpenAI API key
- Configure your news sources
- Define your interests
- Run
npm start
Deploy to Netlify, Vercel, or any static hosting platform. The system auto-commits new content, triggering automatic redeployment.
Cost Estimate
Using GPT-4o-mini, processing 50 articles per day typically costs $1-3 per month in OpenAI API fees. That's it. No subscriptions, no per-user fees, no hidden costs.
Technical Architecture
The system is built as a Node.js application with:
- Modular design: Separate services for fetching, parsing, AI, HTML generation, and git operations
- SQLite database: Tracks processed articles to avoid duplicates
- Error resilience: Continues processing even if individual operations fail
- Comprehensive logging: Tracks every operation for debugging
- Scheduled execution: Runs automatically at configured intervals
- Git integration: Auto-commits and pushes for CD/CI workflows