Intent Recognition

AI Toolkit intent_recognition Starter

Classify user intent using AI

Overview

The Intent Recognition node uses AI to classify user input into predefined intent categories, allowing you to route conversations based on what the user wants.

How It Works

  1. Analyze Input: The AI analyzes the input text (supports Handlebars like {{user_message}})
  2. Classify Intent: Determines which of your defined intents best matches the input
  3. Route: The conversation flows through the output handle matching the recognized intent ID

Output Variable

The node stores the classification result:

Intent: {{intent.intent.label}}
Confidence: {{intent.confidence}}

Configuration Fields

Input handlebars Required

The text input to classify.

Show details

Supports Handlebars syntax:

{{variable_name}}
Possible Intents options Required

Define the possible intents the AI can recognize.

Show details

Each option creates a separate output handle on the node. The flow continues through the handle corresponding to the selected/matched option.

Requirements:

  • Each option must have a unique ID
  • Minimum 2 options required
  • IDs should be valid identifiers (lowercase, underscores)
AI Configuration agent_settings

Configure the AI model and behavior.

Show details

Configuration options:

  • AI model selection
  • Temperature (0.0-2.0) - Higher = more creative, Lower = more consistent
  • Conversation context length - Amount of chat history to include
  • Personality instructions - Defines how the AI interacts

Temperature guide:

  • 0.0-0.5: Focused, deterministic, predictable
  • 0.5-1.0: Balanced between creativity and consistency
  • 1.0-2.0: Creative, varied, exploratory

Output Variables

This node stores its output in a variable that can be referenced in subsequent nodes using the {{variable}} syntax.

{{variable}}.intent object

Detected intent object with id and label

{{variable}}.intent.id string

Intent ID

{{variable}}.intent.label string

Intent label

{{variable}}.confidence number

Confidence score (0-1)

{{variable}}.timestamp string

ISO 8601 timestamp

{{variable}}.error string

Error message (only present on error)

Connection Points

Input
Accepts incoming connections from other nodes
Output
Connects to the next node in the flow

More in AI Toolkit