When most people think of data extraction, they only think of Web Scraping. But what happens when the data you need is trapped inside an Excel-like accounting tool, a legacy ERP system, or a desktop CRM like Slack or Lark?
Data extraction from your desktop is the key to unlocking that information. It acts as the "eyes" of your automation, allowing Octoparse AI to "see" and "read" text, numbers, or statuses directly from any software window on your screen. By converting visual interfaces into structured data, you can build seamless pipelines that connect your local apps to the rest of your digital workflow.
From "On-Screen" to "In-Variable"
Unlike web scraping, which looks at background code (HTML), desktop extraction interacts with exactly what you see. Whether it's a balance in a banking app or a notification in a chat tool, the process follows a reliable logic: Identify → Read → Store.
The core command for this is Get details of UI element in window. Once you capture an element, the bot "reads" its value and saves it into a variable (e.g., OrderNumber). This transforms a static image on your screen into a dynamic piece of data that your bot can "remember" and use in later steps.
Desktop Scraping: Extracting Lists and Tables
The true power of this feature lies in its ability to handle repeating data. If you have a desktop application filled with rows of customer info or transaction logs, you don't need to copy them one by one.
You can build a Desktop Scraper by combining two commands:
Loop through similar UI elements: This tells the bot to find every "row" or "item" that looks the same.
Get details of UI element in window: Inside the loop, the bot extracts the specific data from each row.
This allows you to "scrape" an entire desktop table into a structured list in seconds, ensuring 100% accuracy without the manual grind.
Precision Tools: Anchors and Context
Desktop UIs can be tricky—windows resize, and elements shift. To keep your data extraction "bulletproof," we use two specialized helpers:
Get details of window: This confirms RPA is looking at the correct version of the application before it starts reading.
Get relative element in window: If a piece of data (like a "Price") doesn't have a unique ID, you can tell the bot to "Find the text next to the 'Total' label." This Anchor strategy ensures your extraction stays accurate even if the window layout changes.
Summary: The Value of Desktop Data
Mastering data extraction from your desktop means you are no longer limited by "Export" buttons or complex integrations. You have the power to turn any software into a data source.
By following the Identify → Read → Store cycle, you create a foundation for more intelligent tasks—allowing your bot to not just "click" through your desktop, but to "understand" and act on the information it finds.