← Back to MusicWander

Underground-first

We bias toward niche artists and deep cuts, not viral charts.

Privacy-first

No accounts, no ads, no tracking pixels. Your session stays local.

Fast and lightweight

Vanilla JS, small payloads, and zero app bloat.

How it works

  • Read your Last.fm top artists and tags (3-12 months).
  • Model taste shifts with temporal learning.
  • Explore similar-artist graphs and tag overlap.
  • Generate a queue of 30-second previews.

What you get

  • Discovery, Balanced, and Accessible modes.
  • Hidden gems filters for underground-only picks.
  • Like, skip, and export to build playlists elsewhere.
  • Local history so you can revisit past finds.

Limitations (by design)

  • New Last.fm accounts get less precise matches.
  • Some niche artists have limited preview availability.
  • MusicWander is discovery, not full streaming.

Technology stack

Vanilla JavaScript, Last.fm for data, and iTunes previews for audio. The engine blends collaborative filtering, tag co-occurrence, and similarity scoring.

We keep the stack lean so the app loads fast and runs entirely in the browser.

Created by

MusicWander is an independent project by Guillaume Marical.

Not affiliated with Last.fm, Apple Music, iTunes, or Spotify.

Ready to explore?

Launch MusicWander and start a discovery queue in minutes.

Launch MusicWander