Classifier output types

Classifier outputs follow standardized formats. Understanding this format is essential when working with evaluation sets, webhooks, and API responses.

Classify Output Type

Type Definition

1
type ClassifyOutput = {
2
  id: string;
3
  type: string;
4
  confidence: number;
5
  insights?: Insight[];
6
};

Example

1
{
2
  "id": "invoice",
3
  "type": "invoice",
4
  "confidence": 0.95,
5
  "insights": [
6
    {
7
      "type": "reasoning",
8
      "content": "Document contains itemized charges and payment terms"
9
    }
10
  ]
11
}

Shared Types

Certain types are shared across different run outputs. These provide additional context and information about the classifier’s decisions.

Type Definition

1
type Insight = {
2
  type: "reasoning"; // Currently only reasoning is supported
3
  content: string; // The explanation or reasoning provided by the model
4
};

Example

1
{
2
  "insights": [
3
    {
4
      "type": "reasoning",
5
      "content": "This was classified as an invoice because it contains standard invoice elements including an invoice number, billing details, and itemized charges."
6
    }
7
  ]
8
}

Insights can appear in both extraction and classification outputs to provide transparency into the model’s decision-making process. They are particularly useful when debugging or validating run results.