Virbe Documentation

Call Conv AI

Route the conversation to a configured Conversational Engine (external AI backend) and return its response to the user.

The Call Conv AI node forwards the user's message to a configured Conversational Engine – an external AI backend such as OpenAI Assistant, Dify AI, Dialogflow, or a Custom Endpoint – and streams the response back to the user.

Use this node when you want an external conversational AI system to handle part (or all) of the conversation, rather than using Virbe's native LLM Response or Tool Agent nodes.

Available in: Conversational Flows.


Configuration

FieldDescription
Conversational EngineThe external engine to call. Select from the engines configured under Configurations → Conversational Engines.

How it works

When this node executes:

  1. The current user message and session context are forwarded to the selected Conversational Engine
  2. The engine processes the input using its own AI and logic
  3. The engine's response is returned and streamed to the user as the virtual being's reply
  4. Conversation continues from the engine's response

The node acts as a pass-through – the response content is determined entirely by the external engine, not by the Virbe pipeline. Virbe handles the transport and session context; the engine handles the intelligence.


When to use Call Conv AI

Use this node when:

  • You have an existing AI bot or agent built on another platform (Dialogflow OpenAI Assistant, etc.) and want to route conversations to it from within Virbe
  • You are using a Custom Endpoint to connect a proprietary or self-hosted LLM backend
  • You need the external engine to maintain its own conversation state and memory independently of Virbe's variables and signals

Use LLM Response or Tool Agent when you want to use Virbe's pipeline with a connected AI model's capabilities, as well as your Knowledge Base.


Prerequisites

The Conversational Engine you want to call must be configured and tested before using this node. See Configurations → Conversational Engines for setup guides for each supported engine.

On this page