TypeScript
Next.js
React
Python
PostgreSQL
Prisma
Docker
GitLab CI/CD
Spotify API
Machine Learning
LLM

duudey is a heavy music discovery site that grew out of a web radio project started in 2012.

Today it turns international music coverage into a multilingual, image-forward news wall. It groups related stories from specialist outlets, pairs them with short translated summaries, links each cluster back to its original sources, and keeps the old radio spirit alive through a synchronized Spotify listening module.

The shared listening idea did not disappear. It moved into a Spotify synchronization engine built in-house, where listeners can join an automated room or let a live host take over the queue. The player docked into the current site is the latest version of that older radio instinct: keep the music playing while people browse, read, and discover.

The news side came from a newer frustration. Heavy music coverage is global, fast, fragmented, and usually written in whatever language the outlet happens to publish in. Keeping up means jumping between feeds, tabs, translations, artist pages, and playlists. duudey brings that mess into one place without pretending to replace the original publishers.

The site behaves more like a live music intelligence desk than a traditional blog. Incoming articles are grouped into story clusters, each cluster keeps links to the original outlets, and the reader gets a concise summary in their language before deciding where to click next. The front page surfaces lead stories, fast-moving coverage, deep multi-source stories, tour listings, and genre momentum so the site feels alive even when the same band is being covered from several angles.

The product is intentionally visual. Story cards use publication images when available and fall back through artist imagery when needed, so the experience feels closer to a modern music publication than a database browser. The design leans into a dark editorial wall, sharp pink accents, large headline typography, and dense discovery rails for people who want to wander by scene, artist, or current story.

The technical work sits mostly behind the curtain:

  • A multilingual music intelligence pipeline turns raw music coverage into grouped, summarized story clusters.
  • Artist-aware enrichment connects stories to official Spotify artist data and genre metadata.
  • Event ingestion gives the homepage a tour and concert layer next to the news feed.
  • A synchronized Spotify radio client, descended from the original listening-room project, lets a small group hear the same track at the same time.
  • The site ships as a containerized full-stack application with background workers and automated deployment.

The most important design choice is what duudey does not do. It does not host copied articles, does not hide source links, does not host audio, and does not sell ads against behavioral tracking. It is built as a discovery router for heavy music fans: read the summary, follow the source, jump to the artist, keep listening.

The result is a passion project with a serious engineering spine. It combines editorial aggregation, multilingual processing, music metadata, real-time listening, and a publication-grade interface without exposing the machinery in a way that turns the site into a clone recipe.