It is Pion based, so most of what I’ve written here on Pion and its trajectory applies to LiveKit as well.

Note that it has been adopted by some interesting companies recently.

libwebrtc is mostly a client-side open source media engine from my point of view. You can use it to develop media servers, but it takes more effort and isn't a common approach these days.

No matter what you choose you will need to work about breaking things. It is the nature of software development and a bigger one in WebRTC, where things evolve and break all the time.

Janus is a good place to start here and one that many have used for years with success.

Thanks!

May I know why official libwebrtc isn't mentioned? Does it have difficult learning curve or it's easier to start with one's you mentioned?

I read your "don't expect google to be your outsourcing vendor" and it was a nice read. For my simple use case, maybe at max what I would need in future is to have many to one connections where I'm processing audio/video from multiple users/browsers on one server. Do I have to worry about small details on what I choose?

Janus seems like a good options, so if I work with libwebrtc directly I would have to keep an eye on what's going on and if I'm using some other lib like Pion or Janus people would incorporate the changes by themselves and I don't have to worry about braking anything as I can get support(again it's subject).

Also it would be nice if you can point me to good resource on datachannels, some libs like node-datachannel seem to have binding for it and seems confusing.

Any of the open source media servers here would be able to do that rather easily I’d assume.

Thanks!

I liked it

Thanks David

I think that making an architecture media server agnostic is wishful thinking in the long run.

Different media servers tend to focus and optimize on different things. If you generalize and abstract the architecture you end up losing a lot of the optimizations that might be crucial to media quality of your use case.