Inputs

☐ Prompt / Description (Required)

A clear and concise instruction that describes what this step should do. This is essential for AI to debug and re-find the element if the webpage structure changes.

Good Example: "Extract the product name from the item detail page."

Bad Example: "Try to find and get some text that looks like a product name, but make sure it's not too long and is relevant."

Why it’s bad? Too vague, overly verbose, and doesn’t specify what text to extract.

☐ Element Selector (Required)

Define the element from which text should be extracted using one of the following methods:

CSS Selector

Use standard CSS selectors to identify the element.

XPath

Use an XPath expression to precisely locate the element.

Output Options

☐ Assign to variable (Optional)

Store the extracted text in a variable for later use in the workflow.

☐ Insert to table (Optional)

Save the extracted text into a structured table for further processing.

Usage Example

Scenario: Extracting a product name from an e-commerce page

✅ Good Prompt:

“Extract the product name from the item detail page.”

✅ Good Element Selectors:

CSS Selector: h1.product-title

XPath: //h1[@class='product-title']

✅ Output Options:

  • Assign to variableproduct_name
  • Insert to table → Column: Product Name

Notes

  • The Prompt / Description is required to help AI debug and adjust the selector if the webpage structure changes.
  • Ensure the selector uniquely identifies the target element to avoid extracting unintended text.
  • If the page content is dynamic, consider using a robust selector or referencing workflow data.