3 ms·
Very unlikely, you need the browser for that
by 0-_-0 9mo ago
Very unlikely, you need the browser for that
- petcat 9mo agoI would be very surprised if they didn't still have analytics tracking on the MPEG-DASH streams directly (what yt-dlp is downloading)
- basilikum 9mo agoyt-dlp needs to get the stream from somewhere. It has to fetch the website for that and even execute a JavaScript challenge to retrieve the media endpoint.
- netsharc 9mo agoI'd guess it skips running the JavaScript that reports to the Analytics backend...
- embedding-shape 9mo agoI ended up trying to look it up and found this: > Available options: [...] mark_watched: Mark videos watched (even with --simulate). Only for YouTube https://github.com/yt-dlp/yt-dlp/blob/c8680b65f79cfeb23b342b70ffe1e233902f7933/yt_dlp/YoutubeDL.py#L550 https://github.com/yt-dlp/yt-dlp/blob/c8680b65f79cfeb23b342b... So I'm guessing without that, it doesn't, so would have to add the flag for it to try to count it as a view.
- charcircuit 9mo agoI don't think that flag is related at all.
- embedding-shape 9mo agoWhat else could it mean? Aren't "views" something that matches with "watch"? Maybe I'm missing something from how YouTube works, not a big user, happy to be corrected :)
- charcircuit 9mo agoIt could mean that it marks the video as watched but without increasing the view counter. There are multiple signals that youtube had as a view counter or measuring what parts of a video were watched. Just because youtube knows a video was watched it may not increase the view counter to avoid bots or just because.
- estimator7292 9mo agoGood thing yt-dlp is a browser.
- direwolf20 9mo agoIt's not. It runs enough JavaScript to pass the CAPTCCA (completely automated public turing test to tell computers and computers apart) challenge but it doesn't actually load the page and execute everything.