LiveKit Community

Also, in what format do I get the audio. Is it wav pcm or webm opus?

I know there might be some small latency with the audio, but that's not a problem.

You should be able to do this by using two rooms:
1. publish audio from alice to "staging" room
2. connect to staging room with Go SDK, process audio streams (it's opus)
3. then publish that stream to the actual room

<@U01KZ61GV5Z> thanks for the answer. Additional Q, what if Bob needs to do the same? I need N staging rooms -  where N is the number of participants. Am I right?

Also, what about video? Is it an ok use case for the livekit, or is it hacking too much, and it might end not so well?

You do not need multiple staging rooms.. the backend instance could grab multiple/all audio streams and publish them.

Video from the user could be published to the actual room directly. What this means there will be a bit of potential sync diff between audio and video, depending on your processing time.

And how should I publish the audio from Alice to the staging room? Should I do this serverside?