Sunshine@piefed.ca just in case you haven’t, or someone else needs the tip, you can try both --cookies-from-browser <browser> and --impersonate <browser>
i.e. --cookies-from-browser firefox --impersonate firefox:linux
impersonate requires python-curl_cffi to be installed.
ragebutt@lemmy.dbzer0.com 1 day ago
It’s been ages since I just used yt-dlp so I’m not sure
Tubearchivist has a chrome/firefox plugin for this that makes it a lot easier. You visit youtube while logged in and configure the extension to be able to see your instance of TA, then it just copies your cookie at a set interval.
It’s handy because depending on things your cookie can become stale very quickly. Google does a ton of a/b testing so there’s no “definite” info, I’ve had cookies work for months then all of a sudden need a new one for each download. Others appear to have had similar experiences. Depends on if you’re in the “testing” group ig. Some also suggest that heavy usage can create problems but i have like 8,400 videos in my TA instance and problems seem random
A cookie is generally only required if the video requires you to be logged in though, like a video marked as sensitive or whatever
Another fun issue is proof of origin tokens. These rotate far more often than cookies. Luckily as of now you can generally download videos without them but there are certain things that youtube has already locked behind this. Like if you try to download a video that has subtitles (especially fancy vtt ones with positioning, colors, etc) you’ll often find English doesn’t appear as an option for download unless you supply a pot. If you really get into it there are dockers that can supply pot tokens regularly when a download is initiated.
Google will continue to take steps to lock out people that don’t engage in their ad tracking bullshit. Just means it gets a bit more confusing