Yt-Dlp Sequence Diagrams | Hacker News Reader

Yt-Dlp Sequence Diagrams

app.ilograph.com

105 pointsbillyp-rva32 comments

Auto-generated sequence diagrams are almost always useless. This one is no different.

Strong disagree, I use tools like DeepWiki, Mintlify, Cubic daily to analyze complex codebases. Their diagrams on PRs and generated docs have made it much easier to quickly drop into a complex codebase and understand how it works / make meaningful changes. Especially DeepWiki's diagrams + interactive "Deep Research" / Codemap feature are amazing at building flow diagrams to answer specific questions.

I recently did a complete rewrite of our browser driver codebase by deeply analyzing playwright, puppeteer, selenium, and the chromium source code, relying heavily on auto generated diagrams to decide which parts to read myself / dive deeper on. Yes they occasionally leave things out if the underlying codebase hides the complexity or has no docs, but they're great for thoroughly documented mature codebases.

Why are people upvoting this? I thought I was missing something.

And if they are GenAI they are probably wrong...

"Auto-generated"? As in... a defined translation from one abstraction layer to another? Disagree

A mechanistically generated diagram, i.e. a program reading in something and then emitting a diagram, following a specific algorithm.

> a defined translation from one abstraction layer to another

A solution like the aforementioned cannot reliably present a view with a raised level of abstraction, as that necessitates either making unsound assumptions, or being provided with hints (which can be wrong or expire). Just like with code (because they are code).

And so if this is indeed just a rote visualization like that, sticking to the same level of abstraction (thereby not encountering this issue), then they stipulate it is "almost certainly useless, just like they usually are".

I personally find this agreeable, but then I rarely find diagrams particularly illuminating in general, so...

As in not authored by a human who knows which details are important and which are noise for a given level of abstraction.

Yt-dlp os a powerhouse of the free internet. To break free from walled gardens, I've started to use Seal, an open source F-Droid Android app. It's just a shell over yt-dlp. Unfortunately it started to fail to download some videos. Do you have any tips for better use?

Update the app and yt-dlp.

Just saw that there is an option to update just the lib

now I can download everything again

YTDLnis has been working much better for months now. Switched over from Seal myself.

https://github.com/JunkFood02/Seal

Link to Seal

you can also install it in termux and keep it up to date, install deno for solving challenges etc

> Disclaimer: [...] generated using an LLM and may contain inaccuracies.

Speaks a lot about the HN audience that upvotes these kinds of articles

Half of that audience are now Agents... :-))

F in the chat

How was this created?

Hi, author here. The app itself (Ilograph[0]) is fairly old; its first ShowHN was in 2019.

These particular yt-dlp diagrams were an experiment in getting meaningful sequence diagrams from a code repo using Claude. If you open the black tab on the left, you can see the source of the diagrams. Claude was able to generate most of this, albeit with a good amount of back-and-forth to get the details right, plus some manual tweaking in the end.

[0] https://www.ilograph.com

Kind of despised the way you didn't even let me see a preview of my own created graph without having to signup, and the signup prompt didn't show up until I fixed the mistakes in the YAML. So I pasted something (generated), spent some minutes fixing the mistakes, then get hit with a signup prompt, before being able to see the graph... Just be upfront with the signup requirement, don't let me edit anything, if you're gonna gate it like that.

Thanks for the heads up; that shouldn't have been happening. Please try again.

BTW the year is 2026 and we've got LLMs, is there, at long last, an easy way to have an YouTube downloader I can run in a container (say with Podman) that actually works?

Do I need to be logged in with a Google account for it to work?

I always, since decades I want to say, ran into troubles whenever I download Youtube vids (no matter which tool I use, yt-dlp or other). It works, then doesn't anymore.

Are you simply, reliably, downloading Youtube vids? How? From a VM? From a container? From your main machine while being logged on?

P.S: congrats to the guys at Google, you sure made that hard

You always had to keep whatever youtube video downloader you were using updated at a minimum. For a good while login cookies were only required for downloading restricted videos (whatever youtube considered 18+). Now you need half a web browser to do anything and the update cadence is even crazier.

> You always had to keep whatever youtube video downloader you were using updated at a minimum.

Of course. I always try with the latest versions of yt-dlp.

> For a good while login cookies were only required for downloading restricted videos (whatever youtube considered 18+). Now you need half a web browser to do anything and the update cadence is even crazier.

Yeah at some point in the past I had it mostly working. But basically now, what should I do? Maybe host a VM with an actual browser (not headless: just the real thing) and log into a Google account and then what, use yt-dlp from the same IP?

Basically I don't get it: no clue what needs to be done for YouTube downloads to work reliably (and I'm not talking 18+ stuff: actually I had zero idea there were even vids considered 18+ on YouTube).

>Yeah at some point in the past I had it mostly working. But basically now, what should I do? Maybe host a VM with an actual browser (not headless: just the real thing) and log into a Google account and then what, use yt-dlp from the same IP?

What I've been doing is sshing to my main pc running fedora that has a legit firefox with my google/youtube cookies, have a distrobox with archlinux on it, entering that machine and then running yt-dlp there. Since distrobox has mounts your host /home dir into the guest /home it can easily source cookies from your host pc.

The reason for Archlinux is that it seams to do a much better job than Fedora at keeping yt-dlp updated and with the necessary dependencies to successfully download videos. So this synergy between your main machine (where google cant't tell you are "illegitimate") + a guest machine with the "clandestine" software (in google's view lol) all stitched together by distrobox works well.

Not sure about container but

    yt-dlp https://www.youtube.com/watch?v=xxxxxxxx --cookies-from-browser <your_browser_here>

Works 100% of time for me. Obviously I do not download 10k videos daily so I may just jump over time when it is broken...

Somebody is doing it, given all the ad ridden free YouTube download sites that exist.

Yeah but these are really impossible to use, which is why I tried to host my own, ad-free (and for my own use), yt-dlp wrapped in a OCI container but I never had much success : (

Arrows need to be thicker. "Bold" doesn't work. Nitpick: "bold" doesn't really make sense in opacity settings.

is there a "format" of describing an architetcure/design in such a way that it lens itself to such visualization?

Hey guys, look at what I didn't make. I paid money to have someone/something else do it.

Reply on news.ycombinator.com