Thanks — I’m answering your question about whether to separate dev/staging/prod with Projects or Agents. Use separate Projects for dev, staging, and prod. Projects are the top-level isolation unit: they give independent API keys/secrets, billing/subscription plans, quotas, observability, and service limits. That lets you run dev/staging on lower-cost or free plans, keep usage and logs separate, and prevent test traffic from affecting production. Use Agents inside a Project to run multiple workloads within that environment (multiple agent deployments, different models, etc.). Agents share the Project’s resource pool (session minutes, inference credits, concurrency limits), so putting dev/staging agents in the same Project as prod risks hitting limits and mixing metrics. If helpful, I can show a short example of CLI commands to manage multiple Projects or suggest how to store per-environment keys securely. Which would you prefer? Sources: LiveKit Cloud | LiveKit Docs | Mini tutorial: setting up a LiveKit Cloud project | LiveKit CLI Setup | LiveKit Docs | LiveKit Pricing | Quotas and limits | LiveKit docs