Supported platforms
Extraction runs on yt-dlp, so a link from almost any site it knows will resolve. The sites yt-dlp does not handle are covered by plugins - Spotify is the clearest case, resolved from the public track page rather than by yt-dlp at all.
Plugins are installed by name and are not built in, so Spotify and TikTok need two lines of configuration before they work.
The table below is the shorter list that ArgonFetch tests and supports directly. Anything outside it is worth trying, but is not something we chase when it breaks.
| Platform | Status | Notes |
|---|---|---|
| YouTube | ✅ Video and audio | Muxes separate video/audio streams when no pre-muxed format exists |
| Spotify | ✅ Tracks, playlists and albums | Needs the spotify plugin. Metadata is read from the public pages; audio comes from the matching YouTube Music result |
| TikTok | ✅ | Needs the tiktok plugin |
| SoundCloud | ✅ | Except the licensed catalogue, which is DRM protected and refused - see below |
| ⚠️ Needs a signed-in session | Instagram serves media only to logged-in accounts. Point COOKIES_PATH at an exported session and it works | |
| Playlists | ✅ Any source | Resolved as a collection; download the tracks individually or the whole list as one zip |
Spotify without an API key
Spotify links need the spotify plugin installed, but no credentials. ArgonFetch reads the title, artist and cover from the public track page, then finds the matching result on YouTube Music and streams the audio from there. There is no developer app to register and no secret to rotate.
Playlists and albums work the same way. The track list is read through the same API the Spotify web player uses, which needs no credentials either, and each track is then matched on YouTube Music like a single one.
DRM protected tracks
Some sources hand back media that is DRM protected and cannot be decoded - SoundCloud does this for its licensed catalogue. The link is correct and the track is really there; the source simply refuses to serve it in a usable form, and there is nothing ArgonFetch can do about that.
These answer 415 Unsupported Media Type with the reason in detail, rather than the 404 that would send you looking for a typo. See Errors.
Sites that need an account
Instagram serves media only to a signed-in session, so a link that opens fine in your own browser returns nothing to a server that is not logged in. This is not a bug in the extractor and no retry will help - the request never had the credentials it needed.
Point COOKIES_PATH at a Netscape-format cookies file exported from a signed-in browser and the session travels with the fetch. Without it, the request never had the credentials it needed and no retry will help.
When a platform stops working
Extraction is yt-dlp's job, and sites change. yt-dlp updates itself every 12 hours inside the container, so most breakage fixes itself; restarting the container pulls the current version immediately. If a platform stays broken after that, open an issue.