Today Spotify announced that a number of APIs will no longer be available for new users.
While Spotify won’t immediately take away these endpoints for existing users, it certainly does not inspire confidence for their longevity. Spotify cites “security reasons” as an explanation of why they are closing off these APIs, but we are unclear as to how that will improve security, so we need to assume that Spotify has some other motivations behind this move. More likely than not, they are hatching a strategy to protect their algorithmic assets from data crawlers used by third-party AI companies.
Needless to say, the Spotify services continue to get enshittified, taking away very useful features that developers have come to rely on. ListenBrainz has very different goals, being entirely open-source and part of a non-profit foundation, and we won’t pull the rug out from under our users for monetary or “security” reasons.
On the contrary, our very small team works in direct collaboration with users and developers interested in developing new discovery tools in the music space, and we embrace the variety of ways passionate music lovers want to interact with music collections and recommendations.
Our own frustrations with Spotify’s ever-worsening recommendations was the spark that lit up our interest in recommendations, but again our approach is one of fairness (we don’t tip the scales) focused on the user’s experience rather than the deep pockets of multinational labels.
For developers frustrated that their app stopped working, the good news is that the ListenBrainz team has been working on building some new datasets and API endpoints that offer replacements for what Spotify is taking away. While not everything that Spotify is enshittifying has a direct replacement with ListenBrainz, we can at least offer a path forward for developers.
These features/datasets include:
- Artist similarity*: You can use our ListenBrainz Labs API to explore similar artists.The similarity datasets are still somewhat limited, since we’ve not been able to run all available data through that algorithm, but we plan to do that in the very near future.
- Recommendations for a user: Recommendations are available for ListenBrainz users who send us listen information.
- Custom playlist generation: https://listenbrainz.org/explore/lb-radio/ allows you to customize generated playlists as in-depth as you want (see the documentation for more details)
- Popularity data: Find out the popularity of artists on ListenBrainz, as well as popular tracks for an artist
- Fresh releases: Keep up with new albums coming out, generally or specifically for your taste (as seen on https://listenbrainz.org/explore/fresh-releases/)
Future new datasets include:
- Track similarity
- Album similarity
- Your dream feature here
All of this data is Creative Commons CC0 licensed (read Public Domain) and available on our API endpoints, for free, forever. MetaBrainz is a California 501(c)3 non-profit organization dedicated to creating, maintaining and ensuring that these datasets are available for public use.
And on top of that, the person who coined the term “Enshittification”, Cory Doctorow, has been on our board of directors for 20 years, further ensuring that we’re enshittification proof.
Come play with our data – we’d love your feedback! We’re working hard to make this data better and if it doesn’t yet meet your needs, we hope to meet them soon!
* for the similar artist search, use this value for “algorithm”: session_based_days_7500_session_300_contribution_5_threshold_10_limit_100_filter_True_skip_30
