A terminal music downloader. Feed it a CSV or M3U playlist and it downloads each track, embeds full metadata and cover art.

This script can make mistakes when downloading music: sometimes the scraped site doesn't return the exact song you wanted — it may be unavailable, or a different track with a matching name comes back instead. Testing with 100 songs from my library gave ~11 incorrect downloads. Your mileage will vary. Work will continue to mitigate this, but expect the occasional wrong song.

• Playlist input — point it at a .csv or .m3u/.m3u8 file instead of typing songs by hand.
• Full metadata tagging — title, artist, album, album artist, year, track & disc number, and genre are looked up on Deezer and written into each file.
• Embedded cover art — high-resolution artwork is baked into every FLAC/MP3 (no more loose .jpg files).
• Dinosaur progress bar — see the current song, its phase (searching → downloading → tagging), and overall progress as a 🦖 eats through the bar.
• Resumable — a manifest records finished tracks, so re-running skips what's already downloaded.

• Node.js ≥ 18 — https://nodejs.org
• ffmpeg (for tagging & cover art) — brew install ffmpeg. Without it, downloads still work but tagging is skipped.

git clone https://github.com/super-roomi/Dyno-Music.git
cd Dyno-Music
npm install
npx playwright install chromium

node main.js <playlist.csv | playlist.m3u> [options]

node main.js my-playlist.csv
node main.js favourites.m3u -o ~/Music/Dyno
node main.js --song "Radiohead - Weird Fishes" --visible

CSV — columns title,artist,album (a header row is optional and columns are matched by name when present). A single-column row is read as Artist - Title.

title,artist,album
Weird Fishes,Radiohead,In Rainbows
No One Noticed,The Marías,

M3U — extended playlists using #EXTINF:

#EXTM3U
#EXTINF:257,Radiohead - Weird Fishes
#EXTINF:200,The Marías - No One Noticed

1. Each track is searched on the source site and the top result is downloaded.
2. Metadata is fetched from the Deezer API and matched to the track.
3. ffmpeg embeds the tags and cover art directly into the audio file.

You are responsible for having the rights to any music you download.

The default source (saavn.squid.wtf) is a community mirror and these move or go offline often. If a run fails at the "searching" step with a timeout, the site is probably down or has moved — point Dyno at the current mirror with:

node main.js my-playlist.csv --source https://the-current-mirror.example/
# or
DYNO_SOURCE=https://the-current-mirror.example/ node main.js my-playlist.csv

The scraper expects the same search → download layout as saavn.squid.wtf; a mirror with a different page structure may need the selectors in lib/scraper.js adjusted.

• @super-roomi