My music setup was a pinned browser tab for years. It worked, technically. It also ate RAM doing nothing, my media keys only worked when that tab happened to be focused, and I killed the music at least once a week by closing the wrong window. There’s no official YouTube Music desktop app on Linux, and most of what exists is the same web player in an Electron shell.
So I built my own. It’s called Limusic. Rust and Tauri, no bundled Chromium, audio goes through libmpv.
You sign in with your actual YouTube Music account, so your playlists, likes and subscriptions are just there, and liking a song or adding it to a playlist syncs back like you’d expect.
The part I actually care about is that it’s not bare bones:
- Proper MPRIS, so media keys and your now playing widget just work
- System tray, closing the window doesn’t kill the song, plus play/pause and skip from the tray
- Discord Rich Presence with album art and a live progress bar
- Last.fm scrobbling, connect once and forget about it
- Synced lyrics in a side panel, click a line to jump there
- Listen Together, synced rooms so you and a friend hear the same thing at the same time
- 8 themes, including Catppuccin, because obviously
- Gapless playback, volume normalization, shuffle, repeat, autoplay radio
- Your queue survives a restart
Free and open source, GPL. AppImage (self updating) or an rpm for Fedora, which is what I daily drive. No Flatpak or AUR yet, both are on the list. There’s a Windows build too, and no macOS one yet.
Would really love for people to try it, and I’d genuinely like to hear what everyone thinks. ❤️
Site: https://simohypers.github.io/limusic/ Downloads: https://github.com/SimoHypers/limusic/releases/latest
Not affiliated with Google or YouTube, just a thing I made because I wanted it
I’m always curious about youtube music apps, it seems like there are so many of them everywhere, even f-droid has like 20+ clients listed that all look very similar.
Maybe just a fun app to try making so a lot of people do it?
Certainly looks interesting, but I have to mention that I’m not a fan of the Discord Rich Presence feature. The directory access that this typically requires immediately bumps up any flatpak to a medium risk category.
Granted I’m definitely one of those users that does not care to share my current status to all other users as to what I’m watching/listening to at every moment.
Pretty tired of people posting slop they don’t understand.
This is going to break in a few weeks when YouTube changes again. This is such a waste everyone’s of time. Just noise distracting from people who actually understand the things they build.
Given the first few commits referring to scaffolding phases and the frequent use of em dashes, I’m suspicious this was built with AI assistance. Nothing wrong with that, but a disclaimer of some sort somewhere would go a long way.
First of all, I like rust+tui stack and thanks to sharing to community,but I hate to see open source dev wasting valuable time on reinventing wheel on something constantly change. YouTube’s api, webpage constantly change to prevent third party sw. While we have invidious which can be use as backend, why ppl want to waste their time and abandon project quick? Freetube can use invidious api as source for example. What we need make invidious more backend friendly (i always confused why invidious dev not focus on their api only) and able to have single source of truth on subscription, history across mobile, desktop frontend app
What part, if any, does “AI” have in the development or use of this?
from the gitignore:
/.agents/ /.claude/ /CLAUDE.md /PHASE1-PROMPT.md /RELEASING.md /skills-lock.json /plansso probably quite a bit lmao
and from a brand-new account too. this is slop, people
Does this “dev” even know vibe coded slop can’t be licensed?
According to the U.S. Copyright Office, it absolutely can, but it depends on how much control the user exerts on the final product.
If a single prompt was used and they copy pasted the entire thing, then yes, as you have said, it is not copyrightable.
However, if multiple prompts were used and the end result is compiled/arranged with enough direction by the end user, then it can be copyrighted by compilation.
Generating content with AI is often an initial or intermediate step, and human authorship may be added in the final product. As explained in the AI Registration Guidance, “a human may select or arrange AI-generated material in a sufficiently creative way that ‘the resulting work as a whole constitutes an original work of authorship.’”¹²⁶ A human may also “modify material originally generated by AI technology to such a degree that the modifications meet the standard for copyright protection.”¹²⁷
As several commenters noted, human authors should be able to claim copyright if they select, coordinate, and arrange AI-generated material in a creative way.¹²⁸ This would provide protection for the output as a whole (although not the AI-generated material alone).¹²⁹ A relatively common scenario in registration applications is the combination of human-authored text with AI-generated images. In one early case, for instance, the Office found that the selection and arrangement of AI-generated images with human-authored text in a comic book were protectable as a compilation.
According to who? Was there a legislation or legal case? In which country?
The US copyright office for one
https://www.skadden.com/insights/publications/2025/02/copyright-office-publishes-report
Human authorship is a bedrock of copyrightability, and thus works entirely generated by AI are not copyrightable. (…) Where a work includes both human and AI-generated content, only the human contributions are potentially copyrightable.
That article
completelypractically skips over the section in the report that says that copyright can be applied by compilation to an overall work. See section II - F: “Modifying or Arranging AI-Generated Content”Or you can also just skip to the conclusion of the actual report itself:
As described above, in many circumstances these outputs will be copyrightable in whole or in part—where AI is used as a tool, and where a human has been able to determine the expressive elements they contain. Prompts alone, however, at this stage are unlikely to satisfy those requirements.
https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdfEdit: The article you linked does mention that the works can be copyrighted as a whole, but it does not mention this in the conclusion.
Just because AI was used in the development process doesn’t mean it is bad. Doesn’t mean it’s good either but you can’t just ignore it because AI had something to do with the process. Things being made by humans doesn’t make it good either people can write really bad code.
I think it is the other evidence that it is vibe coded
- Brand new account
- Dev only responds to non-technical or feel-good comments
- other commenters have pointed to AI teltale vibe coding syntax signs in many of the files
- Claude extra files in early commits and changing gitignores to do it.
Advertised Vibe coded projects absolutely are by default bad and here is why:
- Developers do not understand what they have built, so can’t debug it beyond “Claude, fix this”
- Vibe coded projects are easy-come, easy-go very often abandoned (73% of the vibe projects on flathub, and those are the people who thought far enough to decide putting it on flathub was worth it)
- developers can’t answer any questions on security when the app will be handling credentials, be hosted on a server, or attach to some sort of live service -> absolutely gaping security hole that you just have to cross you fingers that Claude didn’t do another “oops, you are completely right to push back,1500 common attack vectors were left open and you are now part of a CSAM botnet, do you want me to fix it?”
If someone wants to make small scripts, one-off test firmware for a home project, writing test cases, design reviews, boilerplates, etc… Are arguably good usecases.
Vibe coding an ongoing software service is almost always a hard no-go and statistically will be abandoned and unupdated, leading to even more security and usability issues.
I think your right it is indeed vibe coded, and vibe code does mean it’s bad by default. If it’s made using AI assistance and a developer that knows what there doing is checking over what the AI writes, then it can be fine. But this person advertised they made it using there very own vibe coded vibe coding software!.
There’s ponytail comments in the source code, which is a Claude code plugin. Definitely AI-written
I would say a great part of this code was at least “AI touched”. There are em dashes in the rust scripts, in the markdowns, in the commit messeges. @Simohypers@lemmy.world - disclose your stuff man! people should know what they are putting on their computers
If you view the first commit, you can see the
CLAUDE.md, so at least in part vibe-coded.Tbh all the vibe-coded apps look the same, and the UI has that vibe-coded style, so I’m not surprised.
AI use does not mean vibecoded
I didn’t say it did, but also “vibe coding” is a fairly new term that does not have a strict consistent definition anyway.
I’m not sure how else to interpret your comment… You found evidence of AI and concluded its at least partly vibe-coded. As for the definition, no I don’t think its all that inconsistent. Feel free to look it up.
Semantics are defined by how a term is used, and people use the term to refer to AI generation of code in any capacity, as well as stricter definitions. The first commit is not just “evidence of AI”, it’s no code and just a markdown file instructing an LLM, which meets at least some definitions of vibe-coding.
Insisting “a word doesn’t mean X, it means Y” doesn’t make it so. Arguing about definitions can be on one of two premises: the word is in practice used to mean Y and not X, or it’s useful to define the word to mean Y and not X. You’ve argued neither, just made an assertion.
it shouldn’t matter.
these tools are here to stay.
They are here to stay, but stuff built with an LLM is statistically way less likely to be maintained for more than a couple months compared to a hand-written thing. Anecdotally I also feel it’s way more likely to have security issues. So I think it’s important to disclose so people have their expectations set correctly.
Subjective, but I’d also trust it less to not have serious mistakes, such as deleting everything in your home directory. Or unexpectedly running insecure HTTP server on all interfaces.
Or perhaps just having a temporary file stored on SSD which gets written to way too much, causing great wear to the drive. Or slow memory leaks.Just things a human would likely think of.
These are all things which are quite likely in human written software as well. Indeed, after years of maintenance these things get ironed out, but most projects do often start with some degree of these problems.
anecdotally, I’ve built a lot of projects and abandoned them because I ran out of time. With an LLM to reduce the input effort I actually think I would’ve maintained them longer.
do you have a source which backs up what you’re claiming?
Ok, so there’s no statistically rigorous review yet, but 73% of LLM-generated apps submitted to FlatHub were abandoned within 6 months. This is about the abandonment rate you should expect to see on any newly announced LLM-written project, and this figure appears to be much higher than hand-written projects (e.g. this study, while not directly comparable, shows a 16% abandonment rate).
The reason for this is sort of obvious to me: an LLM makes it really easy to write code that makes an app work, therefore the threshold of involvement/interest needed to “make an app” is much lower. This in turn removes the selection force that makes software authors likely to be interested in, and thus put in the effort (or tokens) to maintain their software going forward. (this is putting aside the fact that LLMs are still horrible at software architecture and quite bad at infrastructure, making the long-term maintenance more difficult).
I don’t doubt that some people will have a lot of passion for their vibecoded thing, and will keep it going for a while. This might even be the case here given the author went out of their way to post about it on Lemmy and respond to comments, seemingly without an LLM in the loop at that stage of the process. But it’s also important to disclose your LLM use so others can adjust their statistical expectations.
anecdotally, I’ve built a lot of projects and abandoned them because I ran out of time. With an LLM to reduce the input effort I actually think I would’ve maintained them longer.
Likely you would have also built more projects if you had access to an LLM, thus giving you less time to spend maintaining each individual one. Which is kind of the problem here…
The bigger a project gets the more of an issue it becomes too
Fair points, I concede.
As someone who has built a lot of software it’s hard to see people treat LLMs as useless when they are shooting themselves in the foot by doing so.
There will be an interesting time, a few years out. When all venture capital for AI has dried up and AI tokens are 10x as expensive.
All of the passion projects by non-developers who can’t maintain them will dry up.
AI, in the state it exists is a really interesting experiment on what would people make if coding were easily accessible to all. But projects like that, that are against Youtube TOS, which will require massive upkeep will go fallow the first time they make it more difficult to access the data.
A more useful project would be to make a framework to play local files, do lookups though discogs, stuff that won’t change much long term, then modularize the youtube connector so when it fails, the project not useless out of the gate.
This is also screaming bloody murder for a Google cease-and-desist.
Useless? No. Look at the Linux kernel, nothing is written by “AI” but adjacent llm tools are used. It depends entirely on what and how.
If something is vibe coded it can fuck all the way off. If they use a llm to point a human at possible issues (and the human knows what they’re doing) then the worst that happens is the human wastes some time chasing a nonexistent bug.
Ngl I was doing that for a while, google search is ass these days, even 4 years ago, ai was more useful for finding sources than google. I was using bingai at the time and I would also google and try to find sources the regular way, I had way more success just telling ai to only return edu sources (or similar instructions, idr exactly since it’s been a while) People that seek easy results with no effort could’ve got them through chegg back before ai really took off. It was always an active choice to learn how to do a problem rather than just grabbing an answer and faking the proof. (Ngl mostly because exams are proctored or in person, but that hasn’t changed even with AI, students will still fail to passs their classes and get degrees if they rely purely on AI)
Not the person you’re replying to, but
There are no shortages of evidence showing heavy LLM usage causing cognitive decline.
Likewise, if we accept anecdotal evidence, many users report losing understanding of their codebase after letting an LLM do substantial changes for them. Architecture drift is a common problem, too.
LLMs are trained on public code, much of it subpar, so it stands to reason that they are prone to regurgitating security vulnerabilities.
I am not including the plethora of ethical issues with LLMs, which I assume you do not care about coming from a purely utilitarian although short-sighted perspective.
And yet, it does matter.
deleted by creator
Nah I’m with this guy. These anti-AI weirdos drive me nuts. Guy writes you a free tool and all they can do is look a gift horse in the mouth.
Guy writes you a free tool and all they can do is look a gift horse in the mouth.
Misunderstanding.
Main point here was that the person should disclose they used LLMs if they did, and to what degree. So, like, one line of text in the README. Similar to how food products list allergens - might be important to some people.Linus Torvalds also has some vibe-coded stuff for private use that he published. But he also made that clear:
Also note that the python visualizer tool has been basically written by vibe-coding. I know more about analog filters – and that’s not saying much – than I do about python. It started out as my typical “google and do the monkey-see-monkey-do” kind of programming, but then I cut out the middle-man – me – and just used Google Antigravity to do the audio sample visualizer.
I’d like to add my perspective.
I’m not very anti-AI, I use a locally run model to help me write code (especially boring stuff).
For me it’s mostly about “is it worth the time to package this thing in Nixpkgs, add it to my config, and get used to how it works?”
And for vibecoded projects, even if they work now, the answer is “probably not”; more specifically it’s “let’s look if it’s alive in a year, if I’m still interested”.
As such, I really appreciate when people add a disclaimer so that I don’t have to go look at the code only to be disappointed - I also appreciated it back in the day when someone said “this is my first real project” or something, for similar reasons.
And secondly it’s a question of trust. If the author doesn’t disclose such basic things about the way the project has been developed, and in fact avoids answering when asked directly, what other potential problems could there be semi-hidden in the codebase?
yep, that’s basically my thought.
“if you can’t say something nice, don’t say anything at all” is what my momma taught me.
if it’s malicious that’s one thing, if it’s free and open source who gives a rats ass how it was made.
oh no your free thing you didn’t pay for gets abandoned? why don’t you take up the reins?
just feels like asking a lot of people and giving them nothing in return
Linux folks be like: I don’t have a good enough tool for this niche thing. Guess how my next 1000 hours will be spent? 😄
Seriously though, nice work!
not that niche no more bucko
Wow. I’m not a YT music person and MacOS is my daily drive, but this looks incredible. Well done and a very nice contribution to the community.
Go ahead and sign in with your YouTube Gmail credentials, he or she won’t automatically store them for future use.
The source code is open and on Github, and there is also a guide to compiling and running the code locally, if not trusted. I can see that being a paint point though.
But the option of listening to music without signing-in is available. 👍
Ah that’s good. I never sign up to the radio. So basically I don’t wish to login to listen to music.
This is my singular concern. I know it’s open source but I am not knowledgeable nor ambitious enough to scour the code looking for malicious intent. So I would be wondering what assurances we have against this.
Same as my previous answer, the code can be compiled after installing the necessary dependencies and ran locally without the need of trusting my releases. But as a bonus, music can be played and enjoyed without signing in. 👍
You should add this to your OP, it will definitely give you more test pilots. :)
You cannot have it both ways. The source code is open for the sake of transparency. You can’t complain about not trusting the author when the code is readily available. Nothing wrong with laziness, but it’s not that you couldn’t check if you were really that concerned.
The fact that the code is open does not mean it is secure or that you should trust the author. Malware can be in an open source software.
Indeed I’d say it is better to trust a stable project with many contributors rather than a new project just published by a single person.
That does not mean it is not secure, but going through the code and making sure it is secure is a tough job. On a middle sized project it will take you a few days to review it properly, if you even have the knowledge to do that.
Lack of knowledge isn’t laziness. I couldn’t security audit the code even with the right amount of available personal bandwidth.
To what are you referring, though? Are you saying that you won’t ever use FOSS because you do not have the bandwidth to assure it is secure?
I’m referring to the person you replied to, who communicated a lack of ability to audit the code while you honed in expressly on laziness.
You have made a couple leaps in judgment here; I also said nothing about my use or disuse of FOSS projects.
I’m sorry that you feel that way. To my credit, they did indicate a lack of ambition to audit the code. Perhaps saying that’s lazy is a step too far, but I think my point is valid. If you cite your lack of ambition to audit the code and then complain that you cannot trust the code is definitely having it both ways by my book.
You on the other hand cited available bandwidth as the reason to not audit the code. Also a fair reason. You also have the choice not to use the software if you do not trust it. This was why I was trying to understand your comment. If you don’t have available bandwidth to audit the code and that is the reason you don’t care to use it then perhaps that mentality should extend to all FOSS software and you shouldn’t use any of it.
Thank you so much, I really appreciate it.
Best of luck with the YouTube interfaces - anything I have coded to them dies within a year or less due to them changing the rules.
On the flipside: awesome app, probably worth the effort to keep whacking the moles as they come up. Good job so far.
Thank you so much, I’ll take that into consideration, I’m prepared to find ways when the official youtube music changes their ways.
As long as they don’t close their servers for us 😅
Locally stored .mp3 files for the win… all kinds of ways to legally obtain those. You have inspired me to try again to code up my ultimate robo-DJ… we’ll see how far it gets this time.
YouTube, Spotify and other services have a great advantage over locally stored mp3: a strong recommendation system and a virtually infinite library.
It would be good to have a recommendation system which does not depend on these systems, but the difficult part is to get the data to build it.
I reckon the metadata scraped by Anna’s archive may be used to build something like that, but that will eventually become outdated.
I’m kind of over the recommendation system infinite library phase of my music listening. With 10,000+ songs in my library, I’m happy enough to hear an “old favorite” a couple of times a year, and by the time all those old favorites have played, 6+ months has gone by, so…
a recommendation system which does not depend on these systems, but the difficult part is to get the data to build it.
Acousticbrainz / Essentia can be the basis for that, or at least good enough for me to sort out the Metallica from the Enya and play the right mood at the right time from my collection of eclectica.
This looks great and I’ll give it a go… but have you heard of grayjay?
Holy shit I’m bookmarking this for when I get a moment to test !!
If you add support for archive tune’s remote play you could be a great companion app and make my need for spotifies remote play gone!
Damn, dude! This looks amazing! I’ll be trying the Appimage but I’ll be using the flatpak when it’s released.
thank you for sharing. love metrolist and the idea of having a desktop equivalent.
There’s also pear desktop but it’s sort of a web wrapper for the stock yt music experience, and likely has a comparable memory footprint to your situation having it as a pinned browser tab.
Thank you very much, I’m glad other can appreciate aswell. ❤️
Maybe a simple question, but any guides on what this means? I get that it’s supposed to be something in the dev tools, but beyond that, not sure how one is supposed to proceed.
These are instructions to locate the login cookie for YouTube that’s stored in your browser so that you can import it into OPs program. Start with the network tab of the browser’s dev tools.
So how do I do that? How specifically do I find a youtube login cookie? I tried searching, but could not find any useful guides.
When I look in Dev tools in firefox I see a big pile of nonsense. I guess what would be really useful here is an installation guide with actual screenshots and detailed instructions. The instructions provided assume a lot of background knowledge that most people will not have.
Sure I could just log in, but I also don’t want to trust my log in details to a random program I download.
Then you definitely don’t want to give a random app your cookie.
TBF giving an authenticated cookie to a software is much better than login credentials, but still if you don’t trust the software and believe access to your account may lead to sensible data you should not do that.
