Extract Data

The Extract Data node uses AI to extract structured information from unstructured text. Define the fields you want to extract, and the AI will parse the input and return a structured object or array.

Overview

Use the Extract Data node when you need to:

Parse emails, documents, or messages for specific information
Convert unstructured text into structured data
Extract multiple data points from a single input
Build data pipelines that process text content

Configuration

Field	Type	Required	Description
`model`	`string`	Yes	The AI model to use for extraction.
`fields`	`array`	Yes	List of fields to extract from the input.
`isList`	`boolean`	No	When enabled, extracts multiple items as an array. Default: `false`.

Field Configuration

Each field in the fields array has the following properties:

Property	Type	Required	Description
`name`	`string`	Yes	The field name (must be a valid identifier: start with letter/underscore, alphanumeric only).
`type`	`string`	Yes	The data type: `"text"`, `"number"`, or `"boolean"`.
`description`	`string`	Yes	Description of what to extract. Be specific for better accuracy.

Loop Mode

When isList is disabled, the Extract Data node supports Loop Mode:

Field	Type	Default	Description
`loopMode`	`boolean`	`false`	Enable to process 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 executions.
`onError`	`string`	`"stop"`	Error handling: `"stop"` or `"continue"`.

Note: Loop Mode is automatically disabled when isList is enabled, as the node already handles multiple extractions.

Inputs

The Extract Data node accepts inputs from connected upstream nodes. All connected inputs are automatically concatenated and processed together.

Output

Single Object Mode (`isList: false`)

Variable	Type	Description
`result`	`object`	An object containing the extracted fields.

Access individual fields using dot notation: result.fieldName

List Mode (`isList: true`)

Variable	Type	Description
`result`	`array`	An array of objects, each containing the extracted fields.

Example Use Cases

Contact Information Extraction

Fields:

Name	Type	Description
`name`	`text`	The person's full name
`email`	`text`	Email address
`phone`	`text`	Phone number
`company`	`text`	Company or organization name

Input:

Hi, I'm John Smith from Acme Corp. You can reach me at john.smith@acme.com or call 555-123-4567.

Output:

{
  "name": "John Smith",
  "email": "john.smith@acme.com",
  "phone": "555-123-4567",
  "company": "Acme Corp"
}

Order Details Extraction (List Mode)

Fields:

Name	Type	Description
`product`	`text`	Product name
`quantity`	`number`	Number of items ordered
`price`	`number`	Price per unit

Input:

Order contains: 2x Widget Pro ($29.99 each), 5x Basic Widget ($9.99 each), 1x Premium Widget ($49.99)

Output:

[
  { "product": "Widget Pro", "quantity": 2, "price": 29.99 },
  { "product": "Basic Widget", "quantity": 5, "price": 9.99 },
  { "product": "Premium Widget", "quantity": 1, "price": 49.99 }
]

Sentiment Analysis Fields

Fields:

Name	Type	Description
`sentiment`	`text`	Overall sentiment: positive, negative, or neutral
`confidence`	`number`	Confidence score from 0 to 100
`keyPhrases`	`text`	Main topics or phrases mentioned

Best Practices

Write clear field descriptions: The AI uses descriptions to understand what to extract. Be specific about format and content.
Use appropriate data types: Choose number for numeric values, boolean for yes/no fields, and text for everything else.
Valid field names: Field names must start with a letter or underscore and contain only alphanumeric characters (e.g., customerName, order_id).
Use List Mode for multiple items: When extracting multiple similar items (like order line items), enable isList instead of creating separate fields.
Test edge cases: Verify extraction works with variations in input format and missing data.

Extract Data

On this page