### Answer
Here's a one-sentence description for each node type, based on the documentation:
- **Conversation Node**: Use it for pure multi-turn dialogue with the user when no tool calling is needed [(1) ](
) [(2) ](
) .
- **Subagent Node**: Use it when the agent should decide whether and when to call one or more tools during dialogue, based on what the user says [(3) ](
) [(2) ](
) .
- **Function Node**: Use it when a tool must always run deterministically at a fixed point in the flow, regardless of context [(2) ](
) (see also the function node details in [(4) ](
) for custom functions specifically) .
- **Logic Split Node**: Use it to evaluate conditions immediately on entry and branch silently to different destinations, without the agent speaking, especially useful for splitting on dynamic variables or when you don't want to stack too many conditions on previous nodes [(5) ](
) [(2) ](
) .
- **Call Transfer Node**: Use it whenever the agent should transfer the call to a human, another phone number, or a SIP URI — the agent doesn't speak in this node itself, so pair it with a Skip Response conversation node beforehand for a transfer line [(2) ](
) .
- **Extract DV Node**: Use it to extract information from the conversation so far and store it as dynamic variables on entry, rather than for dialogue .
- **End Node**: Use it to end the call at a fixed po