Classifier

The Classifier node uses AI to categorize text into predefined categories. Define your categories with descriptions, and the AI will analyze the input and assign the most appropriate category.

Overview

Use the Classifier node when you need to:

Categorize customer support tickets
Sort emails by type or priority
Route content to appropriate handlers
Tag or label text content automatically

Configuration

Field	Type	Required	Description
`model`	`string`	Yes	The AI model to use for classification.
`categories`	`array`	Yes	List of categories to classify into. Must have at least one category.
`includeJustification`	`boolean`	No	When enabled, returns an explanation for why the category was chosen. Default: `false`.

Category Configuration

Each category in the categories array has the following properties:

Property	Type	Required	Description
`id`	`string`	Yes	Unique identifier for the category (auto-generated).
`name`	`string`	Yes	The category name (e.g., `"bug_report"`, `"feature_request"`).
`description`	`string`	Yes	Description of what content belongs in this category.

Loop Mode

The Classifier node supports Loop Mode for batch classification:

Field	Type	Default	Description
`loopMode`	`boolean`	`false`	Enable to classify each item in an array input separately.
`maxIterations`	`number`	`100`	Maximum number of iterations when loop mode is enabled.
`concurrency`	`number`	`1`	Number of parallel classifications.
`onError`	`string`	`"stop"`	Error handling: `"stop"` or `"continue"`.

Inputs

The Classifier node accepts inputs from connected upstream nodes. All connected inputs are automatically concatenated and analyzed together.

Output

Variable	Type	Description
`category`	`string`	The name of the assigned category.
`justification`	`string`	(Optional) Explanation for the classification. Only present when `includeJustification` is enabled.

Example Use Cases

Support Ticket Classification

Categories:

Name	Description
`bug_report`	User is reporting a bug, error, or unexpected behavior in the product
`feature_request`	User is requesting a new feature or enhancement
`billing_inquiry`	Questions about billing, payments, subscriptions, or pricing
`general_question`	General questions about how to use the product
`account_issue`	Problems with account access, login, or settings

Input:

I've been charged twice for my subscription this month. Can you help me get a refund?

Output:

{
  "category": "billing_inquiry",
  "justification": "The user is reporting a duplicate charge and requesting a refund, which is a billing-related issue."
}

Content Moderation

Categories:

Name	Description
`appropriate`	Content is suitable and follows community guidelines
`spam`	Content is promotional, repetitive, or unsolicited advertising
`offensive`	Content contains inappropriate language or harmful material
`off_topic`	Content is unrelated to the discussion or platform purpose

Email Priority Classification

Categories:

Name	Description
`urgent`	Time-sensitive matters requiring immediate attention
`high`	Important items that should be addressed within 24 hours
`normal`	Regular correspondence with standard response time
`low`	Informational or FYI emails that don't require action

Best Practices

Write detailed category descriptions: The AI relies on descriptions to understand the boundaries between categories. Be specific about what belongs (and doesn't belong) in each category.
Use mutually exclusive categories: Avoid overlap between categories to improve classification accuracy.
Keep category names simple: Use clear, lowercase names with underscores (e.g., feature_request not Feature Request!!!).
Enable justification for debugging: When building or testing your workflow, enable includeJustification to understand why classifications are made.
Consider edge cases: Include categories for unclear or borderline content (e.g., "other" or "needs_review").
Test with diverse inputs: Verify classification works correctly with various input formats and edge cases.

Classifier

On this page