Confirmation on LiveKit Cloud Infrastructure Scalability for Upcoming Launch Event

Getting Started
1

Hi LiveKit Team,

I hope you’re doing well.

We are planning a large launch event and would like to confirm whether our current LiveKit setup can support the expected scale.

Our expected workload is as follows:

  • More than 100 concurrent users.
  • Each user will be connected in a separate one-to-one session.
  • We are currently subscribed to the Ship plan.
  • Our AI Agents are self-hosted, while we use LiveKit Cloud for the real-time infrastructure.

Could you please confirm whether the LiveKit Cloud infrastructure under the Ship plan is capable of handling this workload reliably?

If you need any additional information , we’d be happy to provide it.

We would appreciate your confirmation so we can proceed with confidence in planning our launch event.

Thank you, and we look forward to your response.

2

Hi, congratulations on your success.

With 100 concurrent users each connected in a 1:1 session with an agent, you are talking about 100 “Concurrent agent sessions”. This would exceed the LiveKit cloud-hosted limit of 20 for ship, but since you self-host your agents this won’t be an issue. Though, you should plan for the expected capacity in your agents.

If you use LiveKit Inference, you may hit the concurrent Inference limit (20 on ship).

You should be fine for participant numbers and concurrent connections.

Please also be aware of any Egress limits, if you plan on using that feature, since you get 100 concurrent exports on ship.