This is my experience as well. Compared to for example IRC where you can just open a TCP socket and send ascii as it. I think a more simple protocol would lead to more integrations. It doesn't have to be as simple as IRC. One problem is that implementing audio/video p2p streaming is still super complicated. So we should probably only support text messaging as a first spec. And make separate and isolated audio/video spec!? A good start would be to take WebRTC as is and make a very simple protocol on top of it. And also make use of web push notifications. So the stack already exist and is fairly standardized. We just have to wrap it up in a very simple abstraction/protocol ontop.