HN Reader

Browser extension and local backend that automatically archives YouTube videos

213

123

Cool but ... this also sounds like hording behavior. The number of things I've saved over the years only to throw them away years later and realize that saving them in the first place was a waste of time.

In the 90s my friend's mom would video tape AMC movies. She had 300+ tapes. Maybe she had a few rare ones but now all those movies are available on demand either legally or illegally and in much better quality. Another friend kept all of his 1980s computer magazines (Byte, etc...) and moved these extremely heavy boxes through 30+ years of moves. I doubt he ever opened a single magazine since the moment he saved them. Then they all appeared on The Archive and he finally got rid of them.

To be clear, I have a few youtube videos saved on my local storage. I'm just thinking that saving every video I watch reminds me of the things I've personally over-saved.

Actually that reminds me. I met up with the magazine saving friend recently which is when I verified that he finally got rid of his stash. It made me think about things I'm still saving that if I reflect on I know I will never actually look at. For example I have box of about eight 3.5 inch floppy disks from my Amiga days. The odds that I'm going to get an Amiga or download an Amiga emu and get a drive to read those are close enough to zero that I should throw them away. Similarly I have a book of CD-ROMs of backed up data from the 90s. There's a close to 0% chance that I'm never going to bother look at their contents.

6 months agoby socalgal2

Ive been using Tubearchivist with the extension for this.

https://github.com/tubearchivist/browser-extension

I really like the WebUI of Tubearchivist itself.

6 months agoby erinnh

For YouTube videos I feel are worth archiving, I just add them to playlists on my channel, then periodically download my entire channel using a single yt-dlp command (it can keep track of what's already been downloaded).

6 months agoby frou_dh

> Videos are saved to the ./data/ directory and converted to MOV format using ffmpeg with hardware acceleration

Transcoded (ouch) or just remuxed to a mov container? Have to investigate.

6 months agoby mikae1

Oh, this is huge and important. The number of things I watch that're just gone when I go back to look again!

Youtube is an archive like a grocery store is a food archive. [1]

If it was worth watching in the first place, it's worth saving. Reducing the friction of doing so is going to help a lot of people.

(1: I'm getting this quote wrong, what's the actual and attribution??)

6 months agoby myself248

Speaking of archiving data, is there a search engine that also returns search results from older (no-longer-existent but archived) web pages? The issue I've had with things like the Wayback Machine is that I need to recall the URL or find some other web page still linking to the (now broken) URL.

6 months agoby codethief

I creates something similar in concept but with different goal. I wanted to be able to watch videos with sponsor block on iPad ideally using Plex.

I found self hosted solution like this but I was very dissatisfied with how that worked

on other hand I wanted to check out loco.rs framework, so I decided to implement my own solution.

basically you are able to add channels/playlists on many many platforms that yt-dlp supports, you can select what should be cut out using sponsor block and you choice how many days you want it (videos older that that are automatically deleted)

if you are interested, you can check it out: https://github.com/Szpadel/LocalTube

6 months agoby Szpadel

It started a line of thoughts in me. What if the backend keeps the videos around for a longer period of time, and: * regularly checks youtube, and whenever one archived video gets deleted from youtube, it advertises the video ID on a specific set of Nostr relays * have a different browser extension for yt viewers that activates when the user hits a deleted video. * the backend can stream the video for continuous Bitcoin lightning payment until the stream is kept alive.

So you can make some money on free disk space.

6 months agoby lez

I gave Claude access to supadata YT transcription and obsidian MCP to convert them to "permanent note" format and it's helped tone down my YT addiction a lot

6 months agoby jz10

Interesting. I was looking into creating an extension that manually manipulates and intercepts the vnd.yt-ump [1] requests, then use webcodecs to process everything in the browser.

[1]: https://github.com/gsuberland/UMP_Format/blob/main/UMP_Forma...

6 months agoby computegabe

I don’t really get the purpose of this broadly, because doesn’t YouTube keep videos online unless the creator took them down which is probably not the case 95% of the time? That said for a niche or a high likelihood of a video being removed, or if you really want to be 100% certain it makes sense, but would I be accurate in that statement or am I missing something?

6 months agoby ProofHouse

I'm achieving this with a single yt-dlp script reading url from a clipboard.

6 months agoby ivanjermakov

My archiving app called HEAP can be configured using a simple apple script and yt-dlp to do this too. And since it's a native macOS app instead of a browser extension, it works via all browsers:

https://apps.apple.com/ca/app/heap-website-full-page-image/i...

6 months agoby busymom0

I've had this idea myself so cool to see it implemented.

What I'd really like is a kind of universal web caching backend. So everything I access goes through a cache and I have the option of viewing from cache if something goes offline or changes. I could also mark things as "favourite" so they don't ever expire from the cache. Does such a thing exist?

6 months agoby globular-toast

I've been toying with writing something like this for ages and now I don't have to. Brilliant work, thank you!

6 months agoby hkt

Whoa! I asked about something like this 2 years ago but never got to making anything [1]! Super exciting something like this exists!

[1] https://news.ycombinator.com/item?id=37885584

6 months agoby syntaxing

It would be nice if the extension wrote them to some shared repository. That way, the videos could be preserved for humanity without Google having a say in it.

Added benefit: every video would have to be archived only once.

6 months agoby amelius

Why a backend? What's wrong with just a directory hierarchy?

6 months agoby nottorp

Does this work for videos flagged as age-restricted, or paid subscriber content? IIRC, yt-dlp needs to be passed browser cookies to do so?

6 months agoby dbcooper

Literally a few moments ago I was playing a video from a Nitter instance (Twitter proxy), and wanted to save it with Firefox but nope. Firefox already has the video there, I can see in the developer tools that it's playing from a blob. Even if the blob is then modified on-the-fly, why can't I right-click it and save that blob that you already have? If it's a stream, why can't I dump it? You are already playing it, you already have the bytes, you can "cache" the whole playback into a local file and then let me Right Click -> Save that "cache" file. Had to use `yt-dlp`.

Similar with bookmarks. When I bookmark something, it would be nice if the current state of the page were saved as a fallback (or at least have a checkbox for this behavior). You already have the rendered page right there. I can literally do `Ctrl+S` and save this page as it is right now, even with this half-typed comment. The code for saving the current state of a page is already there. And you can already link binary data to a bookmark (favicon), and text data (description, tags, etc). I wouldn't even mind if I have to give up the description field in exchange for a base64-encoded version of the bookmarked page, I'd be willing to accept that additional storage requirement.[1]

Sites like YouTube that actively work to prevent easy downloads and archiving might still need extensions like the one linked in the submission, but browsers don't make it easy to save some data that they already have, and it's difficult enough to contribute to (or patch for yourself) those massive codebases unless you are already familiar with them.

I'm willing to patch things for myself and repackage them for myself, but unfortunately for these things it's easier and more reliable long-term to just do workarounds.

(Disclaimer: I'm using Firefox.)

[1] EDIT: Having history enabled doesn't seem to do save an offline version, either.

6 months agoby quectophoton

Now add DHT so clients can download videos from each other as a torrent and you solved global video distribution.

6 months agoby WithinReason

https://github.com/ArchiveBox/ArchiveBox

6 months agoby untech

~/os/starchive (main)[56daf7] $ ls -lh data

total 3207312

-rw-r--r-- 1 aa staff 525M Aug 2 09:11 2PMzaym-StM.mov

-rw-r--r-- 1 aa staff 362M Aug 2 09:10 CHbawkGc_os.mov

-rw-r--r-- 1 aa staff 658M Aug 2 09:11 lqR7VV8ftys.mov

~/os/starchive (main)[56daf7] $ ./starachive

Server starting on port 3009...

JSON received: map[videoId:CHbawkGc_os]

Added video CHbawkGc_os to queue. Queue length: 1

Processing video CHbawkGc_os. Remaining in queue: 0

6 months agoby fcpguru