Response Nodes
Response nodes generate the output that the virtual being sends to the user – text, AI-generated replies, or external engine responses.
Generating User-Facing Output
Response nodes can be combined within a single turn – for example, pairing a Text node with a LLM Response node to combine precision with natural conversation flow.
Nodes in This Category
| Node | Description | Available In |
|---|---|---|
| Text | Send a fixed text message to the user – the simplest response node, deterministic and always exact. | Flows, Automations |
| LLM Response | Generate a single-turn AI response using a language model, reading conversation context to produce a natural language reply. | Flows, Automations |
| Tool Agent | Virbe's primary agentic AI node – multi-turn conversations, retrieval-augmented generation, guided task steps, and access to conversation variables and context. | Flows (Agent type only) |
| Call Conv AI | Route the conversation to a configured Conversational Engine (external AI backend) and return its response to the user. | Flows |
| Behavior | Trigger a Metahuman avatar animation or expression – a gesture, emotion, or custom sequence. Limited support: Kiosk only; extendable with custom animation clips via the Unreal SDK. | Flows |
Choosing a Response Node
Use Text for deterministic, scripted replies. Use LLM Response for a single AI-generated turn without ongoing tool use. Use Tool Agent when you need multi-turn reasoning, RAG over the Knowledge Base, or guided steps. Use Call Conv AI if you're routing to an external NLU/dialogue platform instead of Virbe's built-in pipeline.