obsidian says this takes 12 minutes to read.
look at the page updates here.
I like the idea of scrobbling, which is defined as “the publishing of one’s media habits to the internet via software”. This was a big thing back in the day, with last.fm being a very popular way to do this. It allows their users to log and catalogue their songs, favorite or dislike songs, among other things. You could follow people and see what they have been listening to, and you can share these listening habits to anyone who is interested. To be honest, I am missing a lot of history, about last.fm, since it is a website that dates back to 2002, but considering that I only made an account in 2017, and didn’t knew the point of it was, well yeah.
2023 was the start of me finally scrobbling, and I’ve been doing it consistently since then. I like being able to see my musical tastes change as time progresses, seeing myself getting into artists, and then slowly growing out of them as I get somewhat tired of my music. I also like being able to just show my friends what I like to listen to, with a single link to my profile. I just think that it is just neat having all this data that I could use for something, not sure for what, but for something, lmao. However, as time went on, and I started getting worried about privacy, I started having second thoughts about this scrobbing thing, of just having all this data public for all to see. The thing is that, scrobbing is not really private. The point is to show the world what your listening habits are. I didn’t mind random people knowing what music I listen to, they don’t know me, but the issue was more with Last.fm and what they were doing with all that listening data. I felt that they could be selling this to a company, and then this company sells it to another company that can use it to build a profile with and kinda guess things based off what I listen. And sure, they can be like Mr. Goody Two Shoes and promise to never ever sell that data, but considering that (until recently) they have been owned by Paramount, a disgustingly large corporation with probably abhorrent data selling, I decided to jump ship pretty recently, and started using other alternatives.
The two alternatives that I went with are Libre.fm and ListenBrainz. First, Libre.fm is a one guy project that intends to be a FOSS replacement for Last.fm. It is a cool idea, but my god is it pretty limited. It also has not been popular at all, since it lacks a lot of features that Last.fm has (again, one man project). It was pretty close to dying, but it was later brought back to life, and will be active for the time being (look at my profile here). To be honest, I could just delete my Libre.fm account, but I kinda feel sad about it, and I keep it around almost out of pity.
ListenBrainz on the other hand is a much larger project, its feature set is pretty one to one with what Last.fm offers, and it is a really good website, despite how clunky it could be. It is also open source, which is awesome (also the profile here). Also, not a scrobbling service but a tool that has helped me a lot, is Pano Scrobbler, it is an open source app that allows you to scrobble your listens to multiple websites, like Last.fm, Libre.fm, and ListenBrainz, all in a single package. It is nice, lmao. Also, these services allow you to get all the experience of Spotify Wrapped, without the funding of AI powered killer robots, or the helping of the Gestapo. So that is neat, lmao.
As nice as ListenBrainz and Libre.fm are, I always wanted to try hosting my own scrobbing service, as like a backup, or for whenever I want to fully go offline with my scrobbing stuff. I ideally wanted something that would work everywhere, with the apps and programs I had, and wasn’t to complicated to set up. To be honest, I kinda wanted to selfhost a ListenBrainz instance on my PC, but it requires like 60 GB for it to work properly (I assume it needs a bunch of album data from Musicbrainz for it to work properly). I wasn’t going to do that, so welp, Koito. Koito is a selfhosted scrobbling server, and on the surface, it looks pretty nice. It allows you to see listening habits per day, month, year, it allows you to click on a song and see how many times, when, and how much have you been playing it. It is all just . However, it is fucking awful to set up. First of all, it depends on Docker, which immediately makes it awful. Docker is one of those pieces of software where you can follow every single instruction exactly as written, and then you will get some obtuse error that is hard to actually fix. Or the thing that is running inside Docker “works”, but it works in an unacceptable way. This is why it took me 8 hours to get Nextcloud working on Docker, and while I got my compose files set up exactly how I want it, I fear the day the setup breaks and now I have to start up from scratch. But I did manage to get Nextcloud working finally. However, Koito on the other hand? Not so much (at least, in the beginning).
After finally getting it working, Koito lets you import a Listenbrainz backup file as a way to get all your data back, and the way you are supposed to do this is to move the zip file into a specific directory. It does not mention to extract the zip, or anything. It just tells you that it is supposed to update itself when you restart the service. But I did all that, I moved the fucking folder to where it tells me to do so, and I did everything I was asked to. Guess what, nothing happened. I didn’t know if I was doing something wrong. The funny thing? They could’ve added a fucking import data button in the settings, which they of course did not do. I do not know why they decided to make you suffer instead of adding an easy to access button. But here we are.
After a while of going back and forth and shit, and staring at the
logs and then scratching my head to what the fuck was happening, I
realized that the folder that Koito was supposed to use was under
/etc/koito/ but there was nothing there at all. So I went
with the nuclear option. I deleted everything and reinstalled
everything, but this time, I used some path that I can access easily
like /home/user/koito. I saw the folder being created, so I
dragged the zip file with the playback history, into the
imports folder, and then restarted the Docker container.
Nothing happened. It was like, it was completely ignoring the
folder.
I noticed by staring at the logs that it was loading
/etc/koito/ despite me politely asking it to use
/home/user/koito in the compose file I made earlier. No
matter, I decided to instead try doing some configuration with
environment variables, and with
KOITO_CONFIG_DIR_FILE=/home/user/koito I was able to make
Koito read the directory like nothing happened. After all that, I
finally had some sort of music scrobbing server up and running just
fine. So I went to my phone to connect it to Pano Scrobbler, and it gave
me one of the biggest cockblocks of the century: it told me it didn’t
support HTTP links at all. The funny thing is that it seems to be an
Android problem, because the Windows build of the same app was able to
simply connect to the instance, no issues here.
I first tried figuring out NGINX and making it do some reverse
proxying witchcraft to make it use HTTPS, but I failed miserably. Turns
out that it simply didn’t want to work with the settings I wanted to
use. After more searching, I found out that Tailscale has the option to
like, forward certain ports to a different tailnet address (called
services), and this is helpful if you are like me and really not wanting
to deal with Docker and setting up port forwarding that way. So first I
created the service called Koito (named
koito.tailnetaddress.ts.net) and then went into the
terminal and typed
tailscale serve --service=svc:koito --https=4110 localhost:4110,
port 4110 being where it was being hosted on that machine. Afterwards,
you do some additional configuring on Tailscale, and boom, you now
should be able to access Koito on
https://koito.tailnetaddress.ts.net:4110/. Since this is
proper HTTPS provided by Tailscale, this means that Pano Scrobbler
finally was able to accept it with no issues, and I was finally able to
add this to the scrobbler.
Except of course that was not all. I found out that Pano Scrobbler would error out with a 404 when trying to load stuff like the listens list, and I was pretty baffled. I first started off by switching to a different API key, to see if that was the issue, and then when that didn’t worked, I went into the container logs again to see what the issue was. It seems that Koito was able to receive data from Pano Scrobbler with no issues, but when it comes to actually reading back data, it seems that it either is not working, or that feature hasn’t been implemented yet. And sure enough, I decided to look into the issues tab on Github, and found an issue someone filed here, where someone asks for a suggestion for more improvements to the API. From what I can tell, Koito only implements enough of the Listenbrainz API to submit listens to itself, but nothing else. What a shame, tbh.1
At least this was not all for nothing. I now have a local copy of my listen history that I can access at any time, and it works for scrobbling, so yeah. I guess that is all what you need, really. I think the Tailscale setup thing I mentioned before was more exciting to me, because I was finally able to have services like Jellyfin and QBittorrent work with HTTPS, which in the case of Jellyfin, finally gets rid of the annoying error of “oh this is HTTP link wahhhh” that Librewolf often complains about, and for QBittorrent, it seems that I can use Magnet links now, which is very convenient. Going back to Koito however, it is a young project, and it may get some improvements in the future. Or who knows, maybe in the future, when I finally upgrade the storage in my server, I may try selfhosting a full Listenbrainz instance.
tangent 1: kinda sick of the friendslop game trend
To be honest, I am kinda sick of this trend of the friendslop game, aka quickly made games that are intended to be played in multiplayer, they have a unique gimmick which is fun for like two or three weeks and then gets superseded by something else. They are usually priced really cheaply to get players on board, and well, I kinda started getting sick of it.
The thing that gets me is that, I feel the people behind friendslop games often take advantage of FOMO and other stuff like that. I imagine people see a game a couple of their friends are playing, so they go and ask what game it is. They get their answer and they buy it because they are afraid of being left out. They play it for two weeks, and then that friend group move on to the next game, and the last game then proceeds to rot and decay in their Steam library forever. There has been times where I would buy a game in anticipation of being the big thing that my friend groups would play, only for it to just, not happen. So now I have 15 dollars of games that are essentially unplayable because they require a friend group. I could’ve gotten any better game that would’ve provided me hours of fun, but no, I ended up being an idiot, spending it all on these games that I cannot play by myself, and no friends really want to play it with me.
And sure, I guess I can wait it out, but then the next game will come out and now I have to get this other fucking game that will be played only three weeks and then thrown away at the roadside, like a cheap modern TV that barely lasts a year. And there are games that sometimes do last more than three weeks. Lethal Company, PEAK and REPO are three examples of good games that had lasted more than the expected 3 week game cycle, and me and my friends have played it for months.
page updates:
- 2026-06-17: page was created.