We deploy an agent on LiveKit Cloud. Each visitor call is one agent session. Video calls also have a third-party avatar joining the same room. We need 100- 500 agent sessions at once, with LiveKit Inference speech-to-text on those sessions.
For this project, please confirm:
Can concurrent agent sessions (and Inference STT connections) be raised to 100- 500 on Scale, or is that Enterprise only?
Included agent minutes vs $0.01/min overage at that level. On the free plan, included minutes are a hard cap — does that still apply on paid plans?
Cold start and how many agent replicas you’d run for 100-500 concurrent sessions.
Quote and lead time. Planning only, not a purchase.