Skip to main content

​
Where is my agent running and how?

Diagram: you create and configure your agent from the SLNG dashboard or API. The agent runs on SLNG infrastructure and calls Listen, Think, and Speak through the Model Gateway and Execution Layer. Every agent lives in a region you pick. That region is where the actual conversation happens, so it decides where the caller’s audio is handled. In an SLNG Managed region, the agent runs on SLNG’s own infrastructure, in a gateway region across the Americas, Europe, the Middle East, Africa, or Asia-Pacific. In a LiveKit Cloud region, the agent runs on LiveKit Cloud in the US, the EU, or India. You also pick an orchestrator (what glues the models together), either LiveKit or Pipecat. This is the software that powers the conversation, not a place, so either engine can run in any SLNG Managed region. For every agent region code and how an agent’s region is set and changed, see Agent regions. All the models available in SLNG agents are hosted or served by SLNG, and you can add your own provider keys with BYOK. There are different models available per region. You can see the whole list of models here. You manage everything from one place. Whatever region an agent runs in, you create, configure, and launch calls from the same dashboard and API. Only the live conversation happens in the region, and phone lines need to be compatible with the region the agent is in.