It would be easy for them to fingerprint it and block it at a server level, given that it uses some hardcoded headers (which are probably sent in the wrong order versus the browser it's spoofing), doesn't fetch any of the images/stylesheets/etc on the page, and probably fetches scripts/manifests/etc in a predictable order that differs from YouTube's own scripts. Maybe they already do this (fingerprinting and logging, that is), but I haven't heard of anyone being banned for it, so it's probably not something to worry about.
It would also be easy for them to just break the extraction code. The old code used to break every time the signature function changed, and while the current code solves that problem, there are still so many things that they could do to break it, and yet they don't (the current code has only broken a few times that I can think of, and I don't think any of those were intentional on Google's part).
Technically it could violate these parts of the ToS, but they're all grey areas:
- access through anything than the site or 'approved' clients (but youtube-dl does use the site, so it could just be classed as another user agent)
- running automated services against them (running youtube-dl manually is probably fine, even for whole playlists or channels, but running a 'youtube-dl as a service' site like the one in this case is almost certainly not)
- downloading videos (but youtube-dl can also be used for streaming, despite the name)
I'm guessing that Google simply doesn't give a shit, as long as you're not using using it abusively (e.g. offering it as a service or using it to do mass-scraping).
It would also be easy for them to just break the extraction code. The old code used to break every time the signature function changed, and while the current code solves that problem, there are still so many things that they could do to break it, and yet they don't (the current code has only broken a few times that I can think of, and I don't think any of those were intentional on Google's part).
Technically it could violate these parts of the ToS, but they're all grey areas:
- access through anything than the site or 'approved' clients (but youtube-dl does use the site, so it could just be classed as another user agent)
- running automated services against them (running youtube-dl manually is probably fine, even for whole playlists or channels, but running a 'youtube-dl as a service' site like the one in this case is almost certainly not)
- downloading videos (but youtube-dl can also be used for streaming, despite the name)
I'm guessing that Google simply doesn't give a shit, as long as you're not using using it abusively (e.g. offering it as a service or using it to do mass-scraping).
reply