Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
A Youtube video downloader written in Go (github.com/kkdai)
153 points by nadermx on Oct 27, 2020 | hide | past | favorite | 52 comments


As others have pointed out, this is very misleadingly editorialized due to ytdl’s vastly expanded scope. The submission title should be changed to the actual title, “Download Youtube Video in Golang”. (Edit: the title was something like "youtube-dl in golang, actively maintained" when the comment was posted.)

Btw, I should point out there's no shortage of web video downloader projects that support YouTube as well as different, often reduced (compared to ytdl) collections of sites, with different focuses and ergonomics. For instance, two projects I've used/contributed to, both active, YouTube-supporting, and with a focus on Chinese sites:

- https://github.com/soimort/you-get (Python, 8 years old, 35.8k stars)

- https://github.com/iawia002/annie (Golang, 2-3 years old, 12.7k stars)

(Not trying to diminish the project posted; what I'm saying is you can find quite a few actively maintained projects with reduced scope if you look around.)


Enjoy issue 811 in the latter.


That led me to this comment that embodies the modern Chinese war-wolf mentality:

> annie还未活跃到令它们注意的地步,没事

(Rough translation: annie is not yet as active to get noticed, don’t worry)


youtube-dl supports like 600+ websites, not just youtube


`youtube-dl --list-extractors|wc -l`

1163


That's slightly misleading, as it includes stuff like this (just an excerpt, many more options are included)

  youtube
  youtube:search
  youtube:search:date
A more accurate version might be: `youtube-dl --list-extractors | grep -v : | wc -l`

920


1163 is 600+


For me 600+ implies something between 600 and 700. Not almost two times the original number.


Would you apply for a job posting at $600+ per month?


'Actively-maintained' is another way of saying 'We play Whack-a-Mole with YouTube every time they change how the ultimate .mp4 video file url is obtained via the private API'.

ie: If you didn't update youtube-dl that often, you'd often find that it stopped working.


Google and others are adding this obfuscation to commercialize everyone else's content.

The web was originally designed as an open platform.

I would guess most users uploading to Youtube would not want people deterred from watching their content.


Well, their main issue is serving ads with the content.

They don't want applications having the final .mp4 url to embed without them.


I always wondered why they didn't just fetch the necessary parsing-logic plugins from a remote repository at runtime, the first time a given site is requested. There are some security implications, but it seems easier to handle than requiring a full update every time a website rejiggles its pages.


I appreciated it becausecause youtube-twitter didn't rejiggle their entire websites. They deployed obfuscation on a video-by-video basis depending on the content of the video. A new CNN video on wildfires wouldn't get it, but a slightly older video on [insert political topic] would get the newest version of obfuscation (and require an update). This allowed users to identify the particular topics that the tech giants where most interested in suppressing.


Well youtube-dl doesn't do that but there is invidious (currently also unmaintained afaik, it's hard being a scraper in 2020). I guess it would be really easy to add an extractor for invidious in youtube-dl.


Though youtube-dl is taken down in Github it doesn't necessarly mean youtube-dl is going to be unmaintained, or at least the youtube-dl developers are not suddenly disappeared.

What is the authors of youtube-dl's response regarding the takedown?


Their continued silence so far doesn't bode well. Likely they don't want to take RIAA head on and will just let youtube-dl drop and move on with their lives. And that's probably what the RIAA counts on.



The description blurb at the top implies it's not by the original maintainers.

>This is a backup of https://github.com/ytdl-org/youtube-dl since it was taken down by the RIAA.

>I'll wait to see what happens with the GitHub repository and the current maintainer before I do anything with this clone.


yes, looks like it. only member in the organization is https://gitlab.com/iruoy who doesn't seem to be related to the official youtube-dl


Does youtube-dl -U still work?


No.


youtube-dl was already on a path to become unmaintained, or at least too slowly updated to keep up with the fast moving services it targets.

https://web.archive.org/web/20201014194602/https://github.co...


Yeah, I think the takedown effectively kills youtube-dl as it was. It was already a lot of work to keep it up to date, and you'd often get videos you couldn't download and had to wait for an update. The whack-a-mole updating is several full time jobs that no one wants to do: you don't get paid, and might get sued.

I am grateful they did it for so long though!


I suspect the opposite will actually happen. The resulting Streisand effect will be a shot in the arm for youtube-dl or its immediate forks.

A fairly mundane scraping tool is now a fashionable "punk" project that will "stick it to the Man". Mirroring it is already an act of civil disobedience; developing it you can now explore some interesting problems in the space between the email-only world of Linux kernel development and the centralized Github/Gitlab model (i.e. do for issue-tracking what git did for merging). And it's likely simple enough (it's a python scraper, at the end of the day) that it can mobilize large swaths of developers.

Youtube-dl core developers should double-down on it, right now.


>What is the authors of youtube-dl's response regarding the takedown?

All I've seen so far is a blurb on the website https://youtube-dl.org/

>Currently our dev repository is taken down due to DMCA takedown notice by RIAA.

>Downloads still work as usual.


The page doesn't really say... But one of the things which made youtube-dl great was that it worked on almost any website you threw it at.

Looking at the readme and code... This project seems to only support youtube.com? If so, that doesn't really make this a drop-in replacement.


I'd assume it's only YouTube, the usage message has this text [1]:

    Download a video from youtube
The beauty of youtube-dl is that it is so universal, supporting a great number of sites, even the simplest case of giving it a direct link to a video file and it will essentially wget it, and process according to your config.

[1] https://github.com/kkdai/youtube/blob/00a2aefb29c4fbf4fc3bf3...


The youtube-dl software still exists, it's just been taken down from GitHub.


Are the maintainers still maintaining it? The next time Youtube changes something in their "DRM"/obfuscation code, the last version of youtube-dl you got from Github will stop working...


Yep, I also used youtube-dl mainly on non-Youtube websites (Twitter, some news sites, etc). Definitely not a replacement.


The beauty also is that it's integrated with many other projects. You can open URL in smplayer and it will use youtube-dl to get media URL and play the video in that page. The same for mpv, and probably others.


Fantastic. Do we see a beautiful example of Streisand effect with regard to this project?


Might be a good time to sign the latest commits to have some trusted state to revert to... :|


As explained in "Lawful Masses with Leonard French" [1] it is enough to have marketing material which advertised copyright infringement to be taken down.

Don't know if it is enough to claim you can download Youtube videos to fulfill this point. From my point of view it is thin ice these days to claim anything...

[1] https://www.youtube.com/watch?v=wZITscblMBA


The referenced video has a CC-BY license. Downloading and sharing is therefore legal according to copyright law. Anyway, in most countries downloading per se is not illegal when the content wasn't obviously illegaly uploaded. No copyright protection mechanism was circumvented.


At the start of his latest video he indicates the fact that it's not in the readme/end user marketing and just appears in an internal test case, and also that test case appears to be only obtaining metadata and not the video content seems to have changed his mind somewwhat: https://www.youtube.com/watch?v=RCrJM-MrKyI


It's only a cosmetic change from a tech point of view of course, but from non-tech point of view, maybe it would be safer to say it's a youtube offline player than a youtube video downloader.

It would require to pipe downloaded video to vlc or something by default for this to be true, of course.

Agree with other comments here, as well : youtube-dl reference should probably be removed. It will be added back by those who link to the project.


It used to be first rule of youtube-dl was; do not talk about youtube-dl. Now everyone knows. Will they ever learn.


RIAA: I have some very bad news for you.

If I can see a video on my monitor and hear the sound, then I can record my screen and the audio and get a video with 99% of the quality of the original.

Maybe you should ban monitors and speakers next? Or eyes and ears? That would be the only way to stop this.


No, they just need to make a lot of stuff illegal and/or very cumbersome, that's all. And that's definitely possible if enough govts are bribed and lobbied to cooperate with them. They have other corporates already in bed with them. At that point, it becomes a question of going after the few rebels and teaching them a lesson and the rest of the Internet crowd will be left in shock and awe.

Trusted computing 1.0 failed to take off, but you can be sure that it will be reintroduced in future and likely will sneak in under the radar, and trusted computing can very well make all forms of screen recording difficult or impossible except through the analogue hole, and that too can be dealt with in time by IOT devices that will refuse to function unless it can contact ML severs over the Net to "SmartScreen" verify what it is recording... for user security of course.


What a bleak future for humanity.


I can still record my screen with a video camera no matter what


You could. But you probably won't.


Browser extensions would be the next logical choice. Long live the closed source internet! /s

EDIT related: There's a myth about Mozart transcribing Allegri’s Miserere after one hearing, so yes, ears should be DMCA-d as well. And pens. https://www.classicfm.com/composers/mozart/guides/mozart-all...


Yes, if you ignore the actual text of the law anything can be DMCAd. But if you read the law you learn that the scope of 1201 is quite limited.


If youtube-dl is permanently removed from Github, this project will surely follow the same trajectory if it gets any bit popular. So I am not sure what's the point of this.


It's an alternative implementation, and may be useful depending on what happens with the main youtube-dl project.

Perhaps a case of "the more the merrier"?


the lesson of the day is to NOT trust a centralised authority run by microsoft.


Github isn't at fault here, nor is Microsoft. There's little they can do in the face of a DMCA takedown without opening themselves up to a _lot_ of liability. The entire point of that system is to take the hosting provider out of it.


What is it with “X written in Go”?


Why hasn't this project migrated away from github yet?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: