A self-hosted music dashboard that turns your Last.fm scrobbles into intelligently curated YouTube Music playlists - and keeps them that way.
YouTube Music's Replay Mix was my go-to for playing my current favorites - whatever I'd actually been listening to most lately. The problem is how little it really does: hit shuffle and you get one random track, then the rest drop straight back into their original order, there's no way to blacklist a song you're sick of, and there's no insight into how it picks anything. So I rebuilt it properly. This tool reads your Last.fm history, finds the right official upload on YouTube Music for each track (skipping live versions, remixes, nightcore, sped-up, 8D, etc.), weights the result by what you've actually been listening to lately, and keeps the playlist in sync on whatever schedule you set - with every step documented and fully configurable, a real blacklist to drop tracks for good, and a lot more than a single playlist.
On top of that you get weekly snapshot playlists that archive your listening habits, custom tag- and artist-based playlists auto-filled from your Last.fm history, and a full web dashboard with a first-launch setup wizard, built-in scheduler, sync console, history database, webhooks, encrypted backup/restore, and PWA install - so after the initial 5-minute install you never have to touch a terminal or config file again.
The Docker setup includes a web dashboard, a setup wizard, and a built-in scheduler. No terminal needed after install.
Prefer not to use Docker? There's a standalone CLI install too. It runs the same sync engine; you handle scheduling yourself with cron/systemd. The Docker path is recommended for almost everyone.
- Web dashboard to configure everything, browse your playlist, fix wrong matches, and watch syncs run live.
- Smart matching that prefers official Songs over user uploads and rejects nightcore, sped-up, slowed, 8D, etc.
- Recency + play-count weighting so the playlist reflects what you're actually listening to right now.
- Weekly snapshot playlists so you build a long-term archive of how your taste evolves.
- Custom playlists - tag-based (e.g. "Breakcore Mix", "Chill Electronic") auto-filled from your Last.fm tags, artist-based collecting every found song by chosen artists, discovery playlists that recommend songs you've never scrobbled based on your top artists/tracks, or template/filter playlists (top tracks in a window, forgotten favorites, seasonal, and more) built from composable history filters.
- Export any playlist to M3U, CSV, or JSON straight from the dashboard for backups or importing elsewhere.
- Built-in scheduler, webhooks, encrypted backup - and more in the full docs.
Full documentation lives at locko2901.github.io/lastfm-to-ytm.
Ideas on the table:
- ListenBrainz as an alternate scrobble source - a source abstraction behind the fetch layer so the same match/sync engine can run on ListenBrainz history, not just Last.fm. On hold until it can be properly tested (I only use Last.fm) - contributions welcome.
- Spotify as an alternate playlist target - a target abstraction behind the sync layer so the same resolved, recency-weighted track set can be pushed to Spotify playlists, not just YouTube Music. Turns the tool from Last.fm→YTM into a general playlist engine. Contributions welcome.
- ytmusicapi - YouTube Music API wrapper
- Thanks to the Last.fm and YouTube Music communities.
MIT - see LICENSE for details.