HTML Extract

Extract data from HTML using CSS selectors

Overview

The HTML Extract tool parses HTML content and extracts data based on user-defined CSS selector rules. It supports two source modes: (1) JSON mode reads HTML from a specified JSON property on the input item, (2) Binary mode reads HTML from a binary data property. Each extraction rule specifies a CSS selector, a return type (text, html, attribute, or value), and whether to return a single value or an array of all matches. Output is JSON with extracted key-value pairs. Upstream binary data is forwarded unchanged.

Category: Core Nodes
Tool Name: html_extract
Version: 1

Appearance: Icon: lucide-Code | Color: #333377

Node Type

Action (Binary) — handles file/binary data operations

Input / Output

Direction	Port(s)
Input	Input
Output	Output, Error

Credentials

This tool does not require any credentials.

Parameters

Parameter	Type	Required	Default	Description
Source Data	options	No	json	Whether HTML should be read from binary data or a JSON property.
				Options: binary (read HTML from a binary data property), json (read HTML from a JSON property on the item)
Input Binary Field	string	Yes	data	The name of the input binary field containing the HTML file to be extracted. (shown when Source Data is binary)
JSON Property	string	Yes	data	Name of the JSON property containing the HTML to extract data from. The property can contain a string or an array of strings. Supports dot-notation (e.g., “response.body”). (shown when Source Data is json)
Extraction Values	fixedCollection	No	{}	Define one or more extraction rules. Each rule uses a CSS selector to find elements and extracts the specified data.
— Key	string	No	—	The key under which the extracted value should be saved in the output JSON.
— CSS Selector	string	No	—	The CSS selector to match HTML elements (e.g., “h1”, “.price”, “#main a”, “table tr td:nth-child(2)”).
— Return Value	options	No	text	What kind of data should be returned from matched elements.
				Options: attribute (get an attribute value like “class” from an element), html (get the inner HTML the element contains), text (get only the text content of the element), value (get value of an input, select, or textarea element)
— Attribute	string	No	—	The name of the attribute to return the value of (e.g., “href”, “src”, “data-id”). (shown when Return Value is attribute)
— Return Array	boolean	No	false	Whether to return values as an array. If true and multiple elements match, each produces a separate array entry. If false, all matched text is concatenated into a single string.
Options	collection	No	{}	Whitespace handling for extracted values.
— Trim Values	boolean	No	true	Whether to automatically remove spaces and newlines from the beginning and end of extracted values.
Max Concurrency	number	No	10	Maximum number of items to process concurrently.

Output Data

The extracted values are added to the input item’s JSON — each rule writes its result under its own Key, and everything the item already carried, including binary data, comes through unchanged.

{
  "url": "https://example.com/product/42",
  "productName": "Standing desk",
  "prices": ["£349.00", "£299.00"],
  "links": ["https://example.com/a", "https://example.com/b"]
}

• A rule with Return Array on returns one entry per matched element; with it off, the matches are collapsed into a single value.
• A selector that matches nothing produces an empty array with Return Array on, and an empty value with it off — a page whose markup changed will not fail the item, it will quietly return nothing, so validate the values downstream if they are load-bearing.
• A Key that already exists on the item is overwritten.

One item can produce several. When JSON Property points at an array of HTML strings, the node emits one output item per string, each carrying the same input fields plus that string’s extracted values. A single string produces a single item.

Reference the results downstream by expression, e.g. {{ $json.productName }}.

The item fails when the named JSON property does not exist, or when it holds something other than a string or an array of strings.

Usage Examples

• Extract all product prices from an HTML page
• Get the title and meta description from HTML
• Extract all links (href attributes) from a webpage
• Parse an HTML table into structured data
• Extract text content from specific CSS-selected elements

Example Configuration

Pull a heading and every link out of an HTML string on the item:

{
  "type": "html_extract",
  "parameters": {
    "sourceData": "json",
    "dataPropertyName": "htmlContent",
    "extractionValues": {
      "values": [
        {
          "key": "title",
          "cssSelector": "h1",
          "returnValue": "text",
          "returnArray": false
        },
        {
          "key": "links",
          "cssSelector": "a",
          "returnValue": "attribute",
          "attribute": "href",
          "returnArray": true
        }
      ]
    }
  }
}

Extract from an uploaded HTML file, keeping the original whitespace:

{
  "type": "html_extract",
  "parameters": {
    "sourceData": "binary",
    "dataPropertyName": "uploadedFile",
    "maxConcurrency": 5,
    "options": {
      "trimValues": false
    },
    "extractionValues": {
      "values": [
        {
          "key": "productName",
          "cssSelector": ".product-title",
          "returnValue": "text",
          "returnArray": false
        },
        {
          "key": "prices",
          "cssSelector": ".price",
          "returnValue": "text",
          "returnArray": true
        }
      ]
    }
  }
}

Read a meta tag out of a nested response property:

{
  "type": "html_extract",
  "parameters": {
    "sourceData": "json",
    "dataPropertyName": "response.body.content",
    "options": {
      "trimValues": true
    },
    "extractionValues": {
      "values": [
        {
          "key": "description",
          "cssSelector": "meta[name='description']",
          "returnValue": "attribute",
          "attribute": "content",
          "returnArray": false
        }
      ]
    }
  }
}

Scrape several data points from a fetched page:

{
  "type": "html_extract",
  "parameters": {
    "sourceData": "json",
    "dataPropertyName": "pageHtml",
    "maxConcurrency": 10,
    "options": {
      "trimValues": true
    },
    "extractionValues": {
      "values": [
        {
          "key": "pageTitle",
          "cssSelector": "title",
          "returnValue": "text",
          "returnArray": false
        },
        {
          "key": "allHeadings",
          "cssSelector": "h1, h2, h3",
          "returnValue": "text",
          "returnArray": true
        },
        {
          "key": "imageUrls",
          "cssSelector": "img",
          "returnValue": "attribute",
          "attribute": "src",
          "returnArray": true
        }
      ]
    }
  }
}

Analyse an HTML email attachment:

{
  "type": "html_extract",
  "parameters": {
    "sourceData": "binary",
    "dataPropertyName": "emailAttachment",
    "maxConcurrency": 3,
    "options": {
      "trimValues": true
    },
    "extractionValues": {
      "values": [
        {
          "key": "senderName",
          "cssSelector": ".sender-info .name",
          "returnValue": "text",
          "returnArray": false
        },
        {
          "key": "actionLinks",
          "cssSelector": ".cta-button",
          "returnValue": "attribute",
          "attribute": "href",
          "returnArray": true
        }
      ]
    }
  }
}

Error Handling

Mode	Behavior
stop	Halts workflow on first error
continue	Skips failed items, passes successful ones through
errorPort	Routes failed items to Error output port

Tips

Parse HTML from JSON property or binary data and extract values using CSS selectors — outputs structured JSON key-value pairs.

Key Points

1. Data source flexibility — the same field position serves two purposes: Input Binary Field for an uploaded or downloaded file, JSON Property for HTML already sitting on the item.
2. CSS selector power — any valid CSS selector works, including comma-separated groups like h1, h2, h3 and structural selectors like table tr td:nth-child(2).
3. Array vs single values — Return Array decides whether you get every match separately or one combined value.
4. Attribute extraction — set Return Value to attribute and name the attribute to read href, src, content or any data-* attribute.
5. Performance tuning — adjust Max Concurrency to match how many pages arrive at once.

Notes

• This node extracts only. To render HTML, or to turn items into an HTML table, use the HTML node.
• Pair it with HTTP Request for scraping: fetch the page, then point JSON Property at the response body.