Stack: Next.js, Tailwind CSS, PostgreSQL. Content pipeline runs via external automated jobs that scan sources, filter for positive sentiment, generate structured summaries, and publish to the database. The app itself is a clean read-only frontend.
Eight categories: Breakthroughs, Healthcare, Education, Environment, Accessibility, Research, Creative, Business.
Honest limitation: the positive-sentiment filtering isn't perfect yet. Some stories that are more neutral than positive slip through. Still tuning the discovery and classification system.
