To put this into perspective, What.CD [0] was widely considered to be the music library of Alexandria, unparalleled in both its high quality standard and it's depth. What had in the ballpark of a few million torrents when it got raided and shut down. Anna's rip of Spotify includes roughly 186 million unique records. Granted, the tail end is a mixed bag of bot music and whatnot, but the scale is staggering.
True but What.cd had a tremendous amount of notable music not available on Spotify though because it was also sourced from cds, bootlegs, vinyl, tape etc whereas Spotify only includes music explicitly licensed for streaming.
This is true and a category of music that got hit notably hard was live recordings. What had a wide array of live recordings made by sound engineers straight from the mixer. This is something that you simply cannot find now unless you maybe know a guy.
I definitely was not aware Spotify DRM had been cracked to enable downloading at scale like this.
The thing is, this doesn't even seem particularly useful for average consumers/listeners, since Spotify itself is so convenient, and trying to locate individual tracks in massive torrent files of presumably 10,000's of tracks each sounds horrible.
But this does seem like it will be a godsend for researchers working on things like music classification and generation. The only thing is, you can't really publicly admit exactly what dataset you trained/tested on...?
Definitely wondering if this was in response to desire from AI researchers/companies who wanted this stuff. Or if the major record labels already license their entire catalogs for training purposes cheaply enough, so this really is just solely intended as a preservation effort?
> The thing is, this doesn't even seem particularly useful for average consumers/listeners, since Spotify itself is so convenient, and trying to locate individual tracks in massive torrent files of presumably 10,000's of tracks each sounds horrible.
I wouldn’t be so sure. There are already tools to automatically locate and stream pirated TV and movie content automatic and on demand. They’re so common that I had non-technical family members bragging at Thanksgiving about how they bought at box at their local Best Buy that has an app which plays any movie or TV show they want on demand without paying anything. They didn’t understand what was happening, but they said it worked great.
> Definitely wondering if this was in response to desire from AI researchers/companies who wanted this stuff.
The Anna’s archive group is ideologically motivated. They’re definitely not doing this for AI companies.
> The Anna’s archive group is ideologically motivated.
Very interesting, thank you. So using this for AI will just be a side effect.
And good point -- yup, can now definitely imagine apps building an interface to search and download. I guess I just wonder how seeding and bandwidth would work for the long tail of tracks rarely accessed, if people are only ever downloading tiny chunks.
I think the people seeding these are also ideologs and so would be interested in also supporting the obscure stuff, maybe more than the popular. There is no way any casual listeners would go to the quite substantial trouble of using these archives.
Anyone who wants to listen to unlimited free music from a vast catalog with a nice interface can use YouTube/Google Music. If they don't like the ads they can get an ad blocker. Downloading to your own machine works well too.
>> But this does seem like it will be a godsend for researchers working on things like music classification and generation. The only thing is, you can't really publicly admit exactly what dataset you trained/tested on...?
Didn't Meta already publicly admit they trained their current models on pirated content? They're too big to fail. I look forward to my music Slop.
They are too big to fail but they aren’t too big to have to pay out a huge settlement. Facebook annual revenue is about it twice that of the entire global recording industry. The strategy these companies took was probably correct but that calculation included the high risk of ultimately having to pay out down the line. Don’t mistake their current resistance to paying for an internal belief they never will have to.
There are tools that actually download directly from Spotify (needs premium then) but yeah most of them just use the search and download from other sources like YouTube without mentioning it. I won't say which tools download directly out of fear that they get killed but they exist.
Moral and legal discussion aside, this is technically very impressive. I also wouldn’t be surprised if this somehow kickstarts open source music generative AI from China.
Not that we should, but it's technically feasible to have a music streaming server with the torrent as the backend, and selectively download the part of the torrent in respond to on-demand streaming request from the client.
I have Spotify premium but the constant shuffle of content availability has meant I’ve stared routinely archiving my liked songs to avoid any rug pull. Zspotify and co still work a charm.
Both C#m and Db can be played on piano using only the black keys (skipping the 3rd note of the scale). This makes them easy keys for beginners. I'm not sure if that's the reason, but it could be related.
Anecdotally, I know a few vocalists that sound great in these keys and use them as a starting point
Electronic dance music is the biggest genre in the data. So then easy to play shouldn't matter. It's still an interesting question. I think playing Db is pretty nice on the piano even if it's not the easiest.
Monopoly is not a nice thing. Maybe it is convenient, but not nice.
People that gives money to artists are the ones going to concerts and buying music directly to artists. Spotify gives cents to artists, incetivizing awful behaviour (AI music, aggressive marketing, low effort art...).
Some people's urges to destroy all traces of human civilisation astonish me. What do you think Spotify is going to do with all its music when it ceases to exist in however many years? No, we must collectively feed Daniel Ek the Hungry.
Holy crap. This is going to trigger a five-alarm fire at Spotify Engineering. This has got to be among the largest proprietary datasets ever unintentionally publicized by a company.
Am I understanding this wrong? Ripping the metadata I'm fine with. But it sounds like they've ripped every song from Spotify and they're going to release them?
Edit: It seems like they are. Stealing from tens of thousands of artists, big and small, and calling it "preservation" or "archiving" is scummy.
Music piracy is already a thing, not to mention you don't even need to torrent nowadays when music is available for free on YouTube. Those who don't want to pay already don't pay so nothing changes there.
The value of Spotify is the convenience, and this collection does not change that in any way. Your argument would apply if someone were to make a Spotify clone with the same UX using this data.
While I wouldn't call this scummy I do agree with your sentiment. It is technically stealing and those copyrights should be respected.
Full disclosure, I am a career musician AND have been known to pirate material. That said, I think this is a valuable archive to build. There are a lot of recordings that will not endure without some kind of archiving. So while it's not a perfect solution, I do think it has an important role to play in preservation for future generations.
Perhaps it's best to have a light barrier to entry. Something like "Yes, you can listen to these records, but it should be in the spirit of requesting the material for review, and not just as a no-pay alternative to listening on Spotify." Give it just enough friction where people would rather pay the $12/month to use a streaming service.
Also, it's not like streaming services are a lucrative source of income for most artists. I expect the small amount of revenue lost to listeners of Anna's Archive are just (fractions of) a penny in the bucket of any income that a serious artist would stand to make.
It is technically not. Stealing means you have a thing, I steal it, now I have the thing and you do not. You can’t steal a copyright (aside from something like breaking into your stuff and stealing the proof that you hold the copyright), and then a song is downloaded the original copyright holder still have copy.
Calling piracy theft was MPAA/RIAA propaganda. Now people say that piracy is theft without ever even questioning it, so it was quite successful.
In most cases, they couldn't make that decision even if they wanted to. Only independent artists and those that are so large as to have enough sway (Niel Young for example) would be able to. The vast majority of artists you probably listen to don't actually own the rights to their own music.
So let the rights holders make the decision? They would never. Music rights exist for them to extract profit above all else. They don't care about preserving culture or legacy. Which is why it's important that somebody does.
Why is this stealing? You can already listen to everything that's on Spotify with a free account. You are free to also record the audio while it's playing. I suppose grabbing the actual file should't matter? Or is this about releasing? And robbing people of plays they would otherwise get through Spotify?
If you listen to something on Spotify with a free account the artists still get paid. This isn't a case where you're ripping off so mega-corp. You're ripping off thousands of artists from major label ones to tiny indies. Take the metadata and build something cool. Stealing the files and releasing them is something else entirely.
You can record what you play from Spotify and you are already free to play the record again and again and again without the artist being paid.
Most people do not because they find it less convenient than paying 20bucks a month or whatever is the current price in 2025 but that doesn't change the reality.
For most people the appeal of Spotify is not the music itself but the playlists that are shared thanks to its ubiquity. This is the reason other services struggle to make a dent even if they have better quality, UI and algos.
Spotify started by disrupting the market using pirated music by the way so you are pretty much endorsing and encouraging piracy when "paying" your favorite artists through Spotify.
The people I know who go through the trouble of pirating and downloading vast libraries of music are all musicians themselves, or at the very least total music nerds. They don’t want to lose access to their stuff, plus if they ever need to import audio into a DAW, DRM is a no-go. They are the same people who spend large amounts of money on vinyls, and support smaller independent artists through concerts, merch and (back in the day) CDs.
It used to be more mixed, but today, piracy is often the only option to ”own” any media at all.
Unrelated, but I just can't stop myself from saying that I absolutely hate Spotify even though I'm a paying customer. Fuck you Spotify. You were supposed to be a convenient way to discover and listen to music. Now you are only convenient for listening to music, and absolutely terrible for any recommendations. This is sad really. Spotify had good recommendations. It's absolutely in a position where it can provide good recommendations — it has both a vast music library and a vast amount of data on user preferences. And it chooses to push procedural/ai-generated slop instead to earn more money. I thought that maybe buying $SPOT stock will make me more at peace with its greed, but it didn't work. Spotify fucking deserves to crash and burn because it sees paying customers as idiots who might not notice they are fed garbage. Fuck you Spotify, fuck you.
This is more frequent than you would assume. I’ve neither subscribed to Apple Music nor Spotify for this exact reason: I’m a millenial who would like to discover music.
Another extremely annoying effect is, being 40+, they only suggest music for my age. In “New” and “Trending”, I see Muse and Coldplay! I should make myself a fake ID just to discover new music, but that gets creepy very fast.
[0] https://en.wikipedia.org/wiki/What.CD
I definitely was not aware Spotify DRM had been cracked to enable downloading at scale like this.
The thing is, this doesn't even seem particularly useful for average consumers/listeners, since Spotify itself is so convenient, and trying to locate individual tracks in massive torrent files of presumably 10,000's of tracks each sounds horrible.
But this does seem like it will be a godsend for researchers working on things like music classification and generation. The only thing is, you can't really publicly admit exactly what dataset you trained/tested on...?
Definitely wondering if this was in response to desire from AI researchers/companies who wanted this stuff. Or if the major record labels already license their entire catalogs for training purposes cheaply enough, so this really is just solely intended as a preservation effort?
I wouldn’t be so sure. There are already tools to automatically locate and stream pirated TV and movie content automatic and on demand. They’re so common that I had non-technical family members bragging at Thanksgiving about how they bought at box at their local Best Buy that has an app which plays any movie or TV show they want on demand without paying anything. They didn’t understand what was happening, but they said it worked great.
> Definitely wondering if this was in response to desire from AI researchers/companies who wanted this stuff.
The Anna’s archive group is ideologically motivated. They’re definitely not doing this for AI companies.
Very interesting, thank you. So using this for AI will just be a side effect.
And good point -- yup, can now definitely imagine apps building an interface to search and download. I guess I just wonder how seeding and bandwidth would work for the long tail of tracks rarely accessed, if people are only ever downloading tiny chunks.
Anyone who wants to listen to unlimited free music from a vast catalog with a nice interface can use YouTube/Google Music. If they don't like the ads they can get an ad blocker. Downloading to your own machine works well too.
https://en.wikipedia.org/wiki/Useful_idiot
Didn't Meta already publicly admit they trained their current models on pirated content? They're too big to fail. I look forward to my music Slop.
> A while ago, we discovered a way to scrape Spotify at scale.
They wont and shouldn’t divulge the details, but I imagine that would be a fun read!
The data will be released in different stages on our Torrents page:
[X] Metadata (Dec 2025)
[ ] Music files (releasing in order of popularity)
[ ] Additional file metadata (torrent paths and checksums)
[ ] Album art
[ ] .zstdpatch files (to reconstruct original files before we added embedded metadata)
Anecdotally, I know a few vocalists that sound great in these keys and use them as a starting point
People that gives money to artists are the ones going to concerts and buying music directly to artists. Spotify gives cents to artists, incetivizing awful behaviour (AI music, aggressive marketing, low effort art...).
A distributed ripping project to do that would be a fine thing.
Edit: It seems like they are. Stealing from tens of thousands of artists, big and small, and calling it "preservation" or "archiving" is scummy.
The value of Spotify is the convenience, and this collection does not change that in any way. Your argument would apply if someone were to make a Spotify clone with the same UX using this data.
Full disclosure, I am a career musician AND have been known to pirate material. That said, I think this is a valuable archive to build. There are a lot of recordings that will not endure without some kind of archiving. So while it's not a perfect solution, I do think it has an important role to play in preservation for future generations.
Perhaps it's best to have a light barrier to entry. Something like "Yes, you can listen to these records, but it should be in the spirit of requesting the material for review, and not just as a no-pay alternative to listening on Spotify." Give it just enough friction where people would rather pay the $12/month to use a streaming service.
Also, it's not like streaming services are a lucrative source of income for most artists. I expect the small amount of revenue lost to listeners of Anna's Archive are just (fractions of) a penny in the bucket of any income that a serious artist would stand to make.
It is technically not. Stealing means you have a thing, I steal it, now I have the thing and you do not. You can’t steal a copyright (aside from something like breaking into your stuff and stealing the proof that you hold the copyright), and then a song is downloaded the original copyright holder still have copy.
Calling piracy theft was MPAA/RIAA propaganda. Now people say that piracy is theft without ever even questioning it, so it was quite successful.
What’s actually scummy is Spotify paying artists $1 per 1000 streams.
Buy CDs. Use Bandcamp.
So let the rights holders make the decision? They would never. Music rights exist for them to extract profit above all else. They don't care about preserving culture or legacy. Which is why it's important that somebody does.
Most people do not because they find it less convenient than paying 20bucks a month or whatever is the current price in 2025 but that doesn't change the reality.
For most people the appeal of Spotify is not the music itself but the playlists that are shared thanks to its ubiquity. This is the reason other services struggle to make a dent even if they have better quality, UI and algos.
Spotify started by disrupting the market using pirated music by the way so you are pretty much endorsing and encouraging piracy when "paying" your favorite artists through Spotify.
It used to be more mixed, but today, piracy is often the only option to ”own” any media at all.
Another extremely annoying effect is, being 40+, they only suggest music for my age. In “New” and “Trending”, I see Muse and Coldplay! I should make myself a fake ID just to discover new music, but that gets creepy very fast.