Retell Agent Config And Llm

Operator creates an agent and must choose response_engine. They set response_engine.type='retell-llm' but also supply an llm_websocket_url (which belongs to custom-llm).

response_engine.type selects exactly one engine: 'retell-llm' (managed, references a retell_llm_id), 'custom-llm' (your llm_websocket_url), or 'conversation-flow' (a conversation_flow_id). Pick one and supply only its fields — llm_websocket_url is ignored/invalid under retell-llm. Do not mix.

Operator wants the user (caller) to speak first, so the agent stays silent until spoken to. They are unsure whether to omit begin_message or set it to an empty string.

Set begin_message to an empty string to make the agent wait for the user to speak first; a non-empty begin_message makes the agent open the call. Decide deliberately per inbound/outbound direction and verify with a test call which party speaks first.

Integrator wants a managed Retell LLM agent. They call Create Agent first with response_engine referencing an llm_id that does not exist yet.

Create the Retell LLM first (Create Retell LLM → returns llm_id), then create the agent with response_engine={type:'retell-llm', llm_id:<id>}. general_prompt, begin_message, general_tools and model live on the Retell LLM object, not the agent.