Why a proxy is used?

[parasiteoflife]

https://github.com/th-ch/youtube-music/blob/fc52af96b2825c793775a54574265056d5906249/src/plugins/synced-lyrics/providers/YTMusic.ts#L107

This is big privacy concern, our info is sent to your proxy that could be doing whatever you want.

There is also no mention of it in the docs/readme.

Why is it needed?
What does it do?
Where is the source?
Why can't we use one of our own?

[ArjixWasTaken]

Disclaimer: I am not the one who put it there, heck I was not even informed of such change when my PR was merged.

But, it's probably there for korean users, because stuff like lyrics are probably not available there w/o premium.

This is big privacy concern, our info is sent to your proxy that could be doing whatever you want.

If you look at the code, the only info sent is the client and the browseId which is used to obtain the lyrics.

The fetchBrowse method, is the only place where that proxy is used, and it's only used in one place.

Now, I don't know who made that proxy, I don't know if it collects data, but it would only be able to collect what songs you are listening to, which makes sense, how else would it respond with the lyrics.

Is it a privacy concern? Probably. I wouldn't like it if a random proxy knew that my IP address is listening to jpop.

---

Am I in favor of removing the proxy? Yes!
Do I think it was placed there maliciously? No

[JellyBrick]

The /browse endpoint of YTM can only be accessed either by signed in with a dedicated token (not Cookie) or in a signed-out state. Therefore, in countries like South Korea where it’s not possible to use YTM without being signed into a Google account, there’s no way to access the /browse endpoint. This results in an issue where synced lyrics cannot be used in certain countries.

To address this problem, I created and added a API server that simply forwards requests related to synced lyrics via the /browse endpoint. (only the necessary /browse ID and fake client information required for synced lyrics are sent, so any intermediary cannot determine what the browse ID represents)

This proxy does not collect anything, nor is it capable of collecting anything.

[ArjixWasTaken]

Maybe the proxy could be used as a fallback?
We could probably write code that detects if /browse is accessible, and it toggles when to use the proxy or not.

[JellyBrick]

Maybe the proxy could be used as a fallback? We could probably write code that detects if /browse is accessible, and it toggles when to use the proxy or not.

YouTube returns a response with the same structure as a normal response, even when access to the /browse API is blocked, making it difficult to distinguish.

[JellyBrick]

You can find the source code of this proxy at https://github.com/JellyBrick/ytm-browse-proxy