This message was deleted.

Hi Shaan, 1. There are no limits enforced by default, but if you’re running LiveKit Server, you can set some limits by node: https://github.com/livekit/livekit/blob/master/config-sample.yaml#L214 2. We have some benchmarks showing at what point CPU usage becomes saturated by number of publishers/subscribers in a single room: https://docs.livekit.io/oss/benchmark/

this is great thanks!

here’s a few more places where you can set limits: https://github.com/livekit/livekit/blob/master/config-sample.yaml#L125 https://github.com/livekit/livekit/blob/master/config-sample.yaml#L205

to add to this from the client perspective (what @great-umbrella-58130 mentioned are primarily server side limits/config options): The point at which your clients’ browsers will start struggling with both video decoding and 3d rendering highly depends on the capabilities of each client’s machine. A great way to mitigate the need for doing a lot of video encoding is to use livekit’s dynacast feature and for decoding, in case you have 100 people in your room, but only a fraction of those are actually visible in your scene, you can selectively subscribe to the tracks you are interested in and unsubscribe from the other ones.

this is helpful

for server side you could use participant metadata for something like this https://docs.livekit.io/server-sdk-js/classes/RoomServiceClient.html#updateParticipant

beautiful so this is what i am thinking LocalClient writes its position using

updateParticipant

in the metadata property and the server runs the business logic on loop