I could probably get behind what you're saying when I read definitions of "serverless architecture" like this: http://martinfowler.com/articles/serverless.html. Or when it's thrown around talking about containerization (virtualization).
If that's the gripe though, I think it'd be barking up the wrong tree to apply to these types of WebRTC patterns and OP's project.
Dude. You're really focusing on the wrong thing here. No one is lying or attempting to deceive. At worst, the term is being used loosely - at worst. At best it's only referring to the portion of the work that happens during "multi player".
I get your frustration with software development jargon but this seems like an odd choice to attack.
And to two of your points, here's a version that is "serverless", and can run completely from the file system.
But is it truly "serverless"? I mean the file system is serving the file up to the browser!
C'mon.
Oh, and in case you were going to point out that there's a server involved in the link I posted because, "the WebRTC offer/answer exchange is performed manually by the users, for example via IM", (emphasis mine) - the information could just as easily be written on a piece of paper and exchanged via carrier pigeon. The pigeon may still technically be a server, in an architectural sense I suppose, but hopefully at that point you'd agree we're taking the argument past the point of any sensibility.
It'd be cool if you knew what you were talking about. These definitely can be made to be serverless if you really needed to stick to that strict definition of the word. However, there is a small "introduction" broker (server) involved in most of these kind of WebRTC apps. It basically says, "player A - meet player B. player B - meet player A" - after that the clients talk directly to one another and that server can completely go away and the multi-player functionality will work just fine. The clients communicate peer to peer after that.