scdl: the pip release is 2.4.0 and the manifest says 3.0.8
Soundcloud Music Downloader
At a glance
- What is it?
- A SoundCloud downloader that is now a wrapper around yt-dlp with compatibility patches, tagged as unmaintained, with an install guide that lives in somebody else's wiki. Three separate version numbers disagree with each other and two of them are stale.
- Who is it for?
- Take scdl if you want the old interface to a SoundCloud downloader rather than driving yt-dlp yourself, since that is literally what the current version is. Check three things before you install.
- Can I use it commercially?
- Yes, with conditions. GPL-2.0 is a copyleft licence: if you distribute software that includes it, you must release that software's source code under the same licence. Running it internally without distributing it does not trigger that obligation.
- Is it still maintained?
- Yes. The repository last received commits 19 days ago.
- What is it written in?
- Mainly Python, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.
Star history
- scdl-org/scdl4,181 ★ · +58
| Repository | 2026-09-23 | 2026-11-15 |
|---|---|---|
| scdl-org/scdl | 4.1k | 4.2k |
58 Stars in the last 53 days, measured from the daily GitHub snapshot. Compare with other projects
Editorial analysis
Three version numbers, and the oldest one is what you install
The version story is the first thing to sort out, and the page does not try. It says that as of version 3 the script is a wrapper around a general downloader with some defaults and patches for backwards compatibility, which is the whole architecture in one sentence. The repository manifest declares version 3.0.8 and requires Python 3.10 or newer. But the most recent release listed for the package is 2.4.0, published at the end of 2021, and the two releases before that are from 2015, one of them marked as a beta refactor. So there is a two-line history here: a long period of small features, then a rewrite into a wrapper, and a published artefact from before the rewrite. Practically that means installing from a package index gives you the pre-wrapper tool, while the repository and the manifest describe something else entirely. The dependency list confirms the rewrite, since the downloader and a SoundCloud client library are both direct requirements with a recent version floor.
The install guide is hosted in a different account
The installation section of the page is a single link, and the link does not point at this organisation. It points at a wiki under a personal account with a different name from the repository's, on a page named as installation instructions. Everything else on the page, including the issue links in the manifest, points at the current repository. So the one piece of documentation a new user cannot avoid needing is the piece that lives somewhere else, under a name that does not match where you cloned from. That is not necessarily wrong; a long-lived wiki often predates a repository move and gets left behind. But it is exactly the shape of thing that goes stale without anyone noticing, and there is no version or date on it from the visible page, so there is no way to tell from here whether it describes the wrapper version or the pre-wrapper version. The safer path for a new user is the manifest, which lists the entry point, the dependencies, and the Python floor, and to treat the linked wiki as a convenience rather than as the specification.
Six flags cover every shape of link the tool accepts
The command surface is small and organised around what kind of thing you point it at. A link flag takes a URL, and the tool detects on its own whether that is a track, a playlist, or a user, so you do not choose a mode for the link type. A search flag takes a query instead and uses the first result, which is a convenience for pasting a name rather than finding a URL. Then there are four mode flags that say what to do with a user: all tracks including reposts, all uploads without reposts, all favourites, and all tracks the user has commented on. A fifth downloads everything. Two flags control incremental behaviour rather than selection: an archive file plus a flag that downloads only what is not already in it, and a sync mode against the same archive file, which is how you keep a local copy of a playlist that keeps changing. The examples show exactly these combinations, one per shape:
# Download one track
scdl -l https://soundcloud.com/jumpstreetpsy/low-extender
# Download only new tracks from a playlist
scdl -l https://soundcloud.com/pandadub/sets/the-lost-ship --download-archive archive.txt -c
# Sync playlist
scdl -l https://soundcloud.com/pandadub/sets/the-lost-ship --sync archive.txtTag writing is the part that survives the rewrite
What the tool still does itself, rather than delegating, is metadata. It sets tags using a pure-Python tagging library, and the fields it writes are named on the page: title, artist, album, and artwork. That is the feature that keeps this project worth having as a wrapper rather than just calling the underlying downloader directly, because a downloaded file without tags is a file your player cannot group. It also creates playlist files when it downloads a playlist, which is the other piece the downloader does not do for you. The system requirements are correspondingly small and honest: an external media binary for conversion, and Python. So the architecture is a thin layer that adds tagging, playlist files, and an incremental archive on top of something that does the downloading, which is also why the project can say new features are unlikely to be merged when the underlying tool can be asked to do it with an argument.
Configuration is one file, and the test suite reads environment files
Runtime configuration is a single file left in a per-user configuration directory under the tool's own name, which is the convention every one of these Python tools follows and which means the file is easy to find when something behaves oddly. The manifest adds a few details worth knowing. The command is declared as a console script pointing at a function inside the package, so the executable name is the package name and there is no separate entry point to install. The dependency list is four packages, one of which is the option parser, which is why the help output on the page is formatted the way it is. The development group is more interesting: a test runner, a type checker configured over the package and its tests, a linter configured to select nearly every rule and then ignore a long explicit list, a plugin that loads environment files for tests, and a library that reads tags, which is presumably how the tag output is verified. The patches subpackage is excluded from type checking entirely, which is a small sign that the compatibility shims are maintained separately from the logic.
The classifiers and the linter target an older Python than the floor
Two small inconsistencies in the manifest are worth noting because they are the kind of thing that bites a new contributor rather than a user. The project requires Python 3.10 or newer, yet the classifier list includes 3.9, and the linter is configured to target 3.9 as well. So the package metadata advertises support for a Python version its own floor excludes. That combination usually means the floor was raised without updating the adjacent configuration, which is harmless at runtime and mildly confusing in a package index where the classifiers are what a reader scans. The development dependencies include a typing extension package with a marker limiting it to Python versions below 3.11, which is the correct pattern for the same concern in the other direction: someone was paying attention to version-conditional dependencies here. So the inconsistency is narrow, and the marker that fixes it correctly sits two lines above the one that does not.
The project says development is not active, and says what to do instead
The status paragraph is three sentences and they are the most useful thing on the page. It says development is not active, that new features will likely not be merged, especially if they can be covered by passing arguments through to the underlying downloader, and that bug reports and fixes are welcome. That is a maintenance policy stated plainly rather than inferred from an empty commit log, and it tells a prospective contributor exactly where the boundary is. The four dependency floors are recent rather than abandoned, which is the strongest evidence in the repository: a package that nobody maintains does not pin a downloader to a version from this year. So the honest reading is that this is a maintained shim around a moving target rather than an abandoned script, which is a different thing and a different decision. If you need a feature, the project's own advice is to check whether an argument can do it, and if it cannot, the answer is probably no.
Editorial conclusion
Take scdl if you want the old interface to a SoundCloud downloader rather than driving yt-dlp yourself, since that is literally what the current version is. Check three things before you install. The published release and the repository are on different version lines, with the package index behind the code by several years, so pin deliberately and expect the newest features only from source. The project states plainly that development is not active and that feature requests will likely be declined if the underlying downloader can do it already, which makes it a maintenance decision rather than a bet on progress. And the install instructions live in a wiki under a different account from the repository, so read them sceptically.
Frequently asked questions
How can I download SoundCloud songs with metadata?
scdl downloads tracks, playlists, and user libraries from SoundCloud and writes id3 tags afterwards using a pure-Python tagging library, setting title, artist, album, and artwork. It also creates playlist files when downloading a playlist, and needs Python plus an external media binary.
What does scdl do now that it is version 3?
It is a wrapper around a general downloader with some defaults and patches kept for backwards compatibility, so the download itself is delegated. What it still does itself is write the id3 tags, create playlist files, and track an archive file so you can download only new tracks or sync a changing playlist.
Is scdl still being developed?
The project states that development is not active, that new features will likely not be merged especially when the underlying downloader can do them with an argument, and that bug reports and fixes are welcome. Its dependency floors are current, so the shim is maintained against a moving target.
How do I download only new tracks from a scdl playlist?
Pass an archive file with the download-archive option together with the flag that restricts the run to new items. The same archive file is reused with the sync mode to keep a local copy of a playlist that changes over time.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/scdl-org-scdl)Star history chart for your README
[](https://hysenlabs.com/projects/scdl-org-scdl)Was this useful?