Thanks — I’m responding to your question about having separate agent IDs for staging and production and whether the LiveKit CLI can specify an agent name. Short answer: the CLI does not let you set an “agent name” for environment separation. Use separate LiveKit Cloud projects for staging and production so each environment gets its own agent ID, API keys, and subdomain. How to do that: 1. Create two projects (e.g., myapp-staging, myapp-prod). 2. Switch CLI context to the target project: ◦

lk project list

◦

lk project set-default "myapp-staging"

(or

"myapp-prod"

) 1. Deploy the agent in that project (

lk agent create

/

lk agent deploy

). Each project will get a distinct

agent_id

recorded in

livekit.toml

. Notes: • The

agent_name

field is for dispatch routing (to identify agent types like

voice-assistant

), not for separating environments. • Best practice: keep staging and production in separate projects and avoid sharing API keys or agent IDs across them. If you’d like, tell me your project names and which CLI command you’ll run next and I’ll outline the exact commands and what to expect in

livekit.toml

. Sources: Deploying to production | LiveKit Docs | Agent dispatch | LiveKit docs | Agents Framework | livekit/livekit_composite | DeepWiki