The less specific, the more wiggle room for lawyers. You know perfectly well that they don't offer download buttons. I'm actually surprised that ytdl still works, so I guess it's not a serious problem actually and no need to obscure further than they already do. I as an amateur could not figure out how to get the files of yt or soundcloud. Reading scdl.py it looked simple enough but I'd need to register a developer api key. For ytdl I guess the source has become bloated and obscure by now so i wouldn't even try. Having a bunch of dependencies already seemed too cluttered.
Is https://www.youtube.com/watch?v=[videoID] a "video playback page"?
Most "YouTube downloaders", just like any popular web browser, access the Content through these pages.
These pages contain the (escaped) URLs for the video in various formats.
Do the terms restrict the software, e.g. HTTP client, one can use to access to these URLs?
For example, "You agree not to access Content through any software other than Google Chrome, official Google mobile app, etc."
If a user prefers some other http client not approved by Google, is that a violation of the ToS?