# Web browser
<a name="actions-web-browser"></a>

## Overview
<a name="web-browser-overview"></a>

Browser actions provide a comprehensive automation toolkit for performing web interactions. Browser sessions are automatically managed — opened, maintained, and closed as needed. Quick Automate supports flexible element identification through AI-powered natural language descriptions and precise XPath selectors. Key capabilities include intelligent element targeting with fallback mechanisms, self-healing features that automatically handle popups and retry failed actions, data extraction including structured table content, and a full suite of interaction methods from basic clicks and typing to advanced AI-driven multi-step tasks.

**Element identification methods**
+ **Description method:** Uses AI to identify elements using natural language descriptions (e.g., "Log In button"), making automations more resilient to UI changes.
+ **XPath selector method:** Uses precise XPath expressions for exact element targeting (e.g., `//button[text()='Log in']`).
+ **Combined approach:** When both methods are enabled, XPath is tried first before falling back to the description method, combining precision with adaptability.

**Self-Healing features available for type-ins and clicks**
+ **HandlePopups:** Automatically detects and handles unexpected dialogs, cookie notices, and overlays, preventing blocking elements from interrupting workflows.
+ **ValidateAndRetry:** Verifies action success through screenshot comparison and retries if needed, reducing false failures.

## Browser Session Management
<a name="browser-session-management"></a>

Browser instances are automatically created, maintained throughout workflows, and properly cleaned up when complete. The `restart_browser` action allows fresh sessions when needed for scenarios such as login timeouts or session expiration.

## Start Browser Session
<a name="start-browser-session"></a>

Opens a new browser session. Actions inside is scope interact with this browser. The browser closes automatically when finished.

**Properties:**
+ **Action Title** (text-only): The title of the action displayed in the process visualization (default: Browser)

**Note**  
All browser actions must be placed within a browser session scope
Each browser scope is a browser session. Browser automatically closes when actions inside a scope completes.

## Go to Webpage
<a name="go-to-webpage"></a>

Navigates to a specific URL. Used to go to a new webpage in the current browser tab.

**Properties:**
+ **URL** (required): The web address to navigate to (e.g., "example.com") Note: Please provide the complete URL when working with this.

## Click
<a name="click"></a>

Clicks on a webpage element. Used to interact with buttons, links, or other clickable elements.

**Properties:**
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description of the browser element to click (e.g., "Submit button"). Don't use the action in the element, just the description of the element
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Mouse Button** (dropdown): Which mouse button to click - Left, Middle, or Right (default: Left, optional)
+ **Click Type** (dropdown): Single or Double click (default: Single, optional)
+ **Get File Download** (radio button): Enable if click downloads a file (default: OFF, optional)
+ **Downloaded File** (output, conditionally required): Variable name storing the downloaded file (default: downloaded\$1file)

**Element Identification:**
+ At least one element identification method must be enabled
+ When both methods are enabled, XPath selector is tried first before falling back to description

## Enter Text
<a name="enter-text"></a>

Types text into an input field. Used to fill in forms, search boxes, or other text inputs on webpages.

**Properties:**
+ **Text to Enter** (required): The text you want to type into the field (e.g., "Order \$112345")
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Order number field")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Replace Existing Text** (checkbox): Clear any pre-existing text in the field before typing (default: ON)

**Element Identification:**
+ At least one element identification method must be enabled. Recommended to use element description and move to element selector only if required.

## Enter Credential
<a name="enter-credential"></a>

Types a username or password. Used to securely sign in to a website using saved credentials.

**Properties:**
+ **Credential** (dropdown, required): Choose which saved credential to use from automation group
+ **Value to Enter** (dropdown, required): Choose whether to enter Username or Password (default: Username)
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Username field")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='username']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)

**Element Identification:**
+ At least one element identification method must be enabled

## Select from Dropdown
<a name="select-from-dropdown"></a>

Chooses a value from a dropdown. Used to make selections in forms, filters, or other dropdown menus on webpages.

**Properties:**
+ **Value to Select** (required): The option you want to select from the dropdown field (e.g., "Complete")
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Order status dropdown")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//select[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)

**Element Identification:**
+ At least one element identification method must be enabled

## Extract Text
<a name="extract-text"></a>

Reads a value from a webpage. Used to capture the text from an individual field.

**Properties:**
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Total amount field")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Extracted Text** (output): Variable name storing the extracted text (default: extracted\$1text)

**Element Identification:**
+ At least one element identification method must be enabled

## Extract Table
<a name="extract-table"></a>

Reads data from a webpage table. Used to capture structured information found in tables.

Intelligently extracts structured data from web pages using natural language targeting or XPath selectors. Features include AI-enhanced extraction, multiple attribute extraction (text, links, etc.), automatic column generation for attributes, and data table output for programmatic manipulation.

**Properties:**
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Order details table")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//table[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Extract Multiple Pages** (radio button): Continue extracting data from additional pages if they exist (default: OFF, optional)
+ **Next Page Button Selector** (conditionally displayed): XPath expression for next page button (e.g., "//\$1[@id='nextButton']", optional)
+ **Intelligent Data Extraction** (checkbox): Use AI to extract and structure data from complex tables (default: ON). Do not use for large table sizes of 100 rows or more. This option should be used for complex tables or table-like structures, where normal table extraction from a web browser fails
+ **Data to Extract** (multi-select dropdown): Choose what to extract - Text, Links (Href), or Element selectors (Xpath) (default: Text, optional)
+ **Extracted Table** (output): Variable name storing the extracted data table (default: extracted\$1table)

**Element Identification:**
+ At least one element identification method must be enabled

**Usage Notes:**
+ Additional data types are saved in separate columns (e.g., columnName\$1Href for Links)
+ Disable intelligent extraction for tables with 100\$1 rows for better performance

## Take Screenshot
<a name="take-screenshot"></a>

Captures an image of the webpage. The screenshot is saved to a file to be used later in your process.

**Properties:**
+ **File Name** (optional): Name for the saved image file. Auto-generated if empty (e.g., "Screenshot image")
+ **File Type** (dropdown): Choose the image format - PNG or JPEG (default: PNG, optional)
+ **Screenshot File** (output): Variable name storing the screenshot file (default: screenshot\$1file)

## Visual Q&A
<a name="visual-qa"></a>

Answer questions about a webpage. Uses AI to visually analyze the page and answer true/false questions about the content.

**Properties:**
+ **Question** (required): Enter your true/false question about the webpage content (e.g., "Is the product in-stock?")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Answer** (output): Variable name storing the true/false response (default: is\$1true)

## Click and Upload Files
<a name="click-and-upload-files"></a>

Uploads files to a webpage. Used to automate file pickers by clicking on a file upload button and choosing the files to upload.

**Properties:**
+ **Files to Upload** (required): The list of files to be uploaded, typically stored in a variable (e.g., [my\$1file1, my\$1file2])
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Upload button")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Mouse Button** (dropdown): Which mouse button to click - Left, Middle, or Right (default: Left, optional)
+ **Click Type** (dropdown): Single or Double click (default: Single, optional)

**File Requirements:**
+ Must be an array of media/file objects

**Element Identification:**
+ At least one element identification method must be enabled

## Enter Keystroke
<a name="enter-keystroke"></a>

Sends a key command or shortcut. Used to automate typing individual keys or key combinations on a webpage.

**Properties:**
+ **Special Keys** (dropdown, required): Select modifiers like Ctrl, Alt, Shift or singular keystrokes like Enter, Tab, Pagedown
+ **Command Key** (optional): Enter the key to be used in combination with the Special key (e.g., "c" for Ctrl\$1c)
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Search bar")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)

**Element Identification:**
+ Both element toggles can be OFF - keystroke is sent to the current active field if no element is defined

## Wait for Element
<a name="wait-for-element"></a>

Checks for an element to appear. Used to wait for changes or validate conditions on a webpage. Returns true if the element is found within the maximum wait time, false if not.

**Properties:**
+ **Action Title** (text-only): Display name for the action
+ **Element Selector** (required): XPath expression to define the element you want to check (e.g., "//input[@id='xyz']")
+ **Max Wait Time** (optional): Maximum time to wait in seconds for the element to appear (default: 30)
+ **Element Found** (output): Variable name storing whether the element was found (True/False) (default: is\$1found)

**Validation:**
+ Max wait time must be an integer or decimal

## Find Matching Elements
<a name="find-matching-elements"></a>

Gets elements matching a pattern. Used to get a list of similar elements on a webpage that you can process one by one.

**Properties:**
+ **Element Selector** (required): XPath expression to define the pattern for elements you want to get (e.g., "//div[@class='abc']")
+ **List of Elements** (output): Variable name storing the list of found elements (default: element\$1list)

**Usage Notes:**
+ The list is empty if no matching elements are found
+ Use for processing multiple similar elements like search results or hyperlinks

## Wait for New Window
<a name="wait-for-new-window"></a>

Checks for a new window to open. Used after clicking a link or button that opens a new child window in the current browser session.

**Properties:**
+ **Max Wait Time** (optional): Maximum time to wait in seconds for the new window to appear (default: 30)
+ **Window Identifier** (output): Variable name storing the identifier for the new window (default: window\$1id)

**Usage Notes:**
+ If time exceeds and window not found, an exception occurs
+ Window identifier can be used to switch focus or close the window

**Validation:**
+ Max wait time must be an integer or decimal

## Switch Window
<a name="switch-window"></a>

Changes browser windows. After switching focus to the specified window, subsequent browser actions interact with that window.

**Properties:**
+ **Window Identifier** (optional): The window you want to switch to, typically stored in a variable (e.g., window\$1id). Leave empty to switch to the main window

## Close Window
<a name="close-window"></a>

Closes a browser window. Once closed, the automation returns to the main window of the current browser session.

**Properties:**
+ **Window Identifier** (required): The window you want to close, typically stored in a variable (e.g., window\$1id)

## Save to Clipboard
<a name="save-to-clipboard"></a>

Copies text to the clipboard. Used to save text that you can paste later on. Existing clipboard text is replaced.

**Properties:**
+ **Text to Save** (required): The text you want to copy to the clipboard (e.g., "Order \$112345")

## Paste from Clipboard
<a name="paste-from-clipboard"></a>

Inserts text from the clipboard. Used to paste copied text into webpage fields.

**Properties:**
+ **Find Element with Description** (radio button): Use natural language description to identify the element with AI (default: ON)
  + **Description Field** (conditionally required): Natural language description (e.g., "Order number field")
+ **Find Element with Selector** (radio button): Use XPath expression to identify the element (default: OFF)
  + **XPath Field** (conditionally required): XPath expression (e.g., "//input[@id='xyz']")
+ **Self-healing** (checkbox): Use AI to handle unexpected popups automatically (default: OFF, optional)
+ **Replace Existing Text** (checkbox): Clear any pre-existing text in the field before pasting (default: ON)

**Element Identification:**
+ At least one element identification method must be enabled

## Refresh Webpage
<a name="refresh-webpage"></a>

Reloads the current webpage. Some pages may redirect on refresh.

**Properties:**
+ **Action Title** (text-only): Display name for the action

## Restart Browser
<a name="restart-browser"></a>

Closes and reopens the browser. A new session starts from a blank page.

**Properties:**
+ **Restart All** (checkbox): Close and restart all browser windows, not just the current one (default: OFF, optional)

## Limitations
<a name="web-browser-limitations"></a>

**Element Identification Uncertainty**: If the AI cannot reliably identify UI elements, actions fail with 'LowConfidence' errors. In these cases, use more specific XPath selectors or element descriptions. **Browser Compatibility**: The system operates exclusively with Chrome browser instances and utilizes Playwright as the underlying automation framework, which means compatibility is limited to Chrome-supported web technologies and may not work with browser-specific features from other vendors. **Dynamic Content Challenges**: Highly dynamic pages with frequent layout changes, heavy JavaScript frameworks, or complex single-page applications may require additional wait times, retry logic, or specialized handling approaches. **Performance Overhead**: AI-powered features like natural language element targeting and self-healing capabilities require more processing time than direct XPath targeting.