Interface WebDriver

  • All Implemented Interfaces:
    java.io.Closeable , java.lang.AutoCloseable

    
    public interface WebDriver
     implements Closeable
                        

    WebDriver defines a concise interface to visit and manipulate webpages. @mcp

    The webpage is rendered to a Document Object Model (DOM) in a real browser, and the interface provides methods to control the browser, select textContent and attributes of Elements, and interact with the webpage.

    All actions and behaviors are optimized to mimic real people as closely as possible, such as scrolling, clicking, typing text, dragging and dropping, etc.

    The term document here refers to a Document Object Model (DOM) within a browser.

    The methods in this interface fall into three categories:

    • Control of the browser itself

    • Selection of textContent and attributes of Elements

    • Interact with the webpage

    Key methods:

    • navigate: navigate to a URL.

    • currentUrl: get the current URL displayed in the address bar.

    • scrollDown: scroll down on a webpage to fully load the page. Most modern webpages support lazy loading using ajax tech, where the page content only starts to load when it is scrolled into view.

    • pageSource: retrieve the source code of the webpage.

    For each document, there are several properties that represent the URL of the document:

    • driver.currentUrl(): Returns the URL displayed in the address bar, it can be either navigated or not.

    • driver.url(), document.URL: Returns the URL of the document.

    • driver.documentURI(), document.documentURI: Returns the URI of the document.

    • driver.baseURI(), document.baseURI: Returns the base URI of the document.

    • document.location: Represents the location (URL) of the current page and allows you to manipulate the URL.

    In the Document Object Model (DOM), the relationship between document.URL, document.documentURI, document.location, and the URL displayed in the browser's address bar is as follows:

    • driver.currentUrl():

    • driver.url(), document.URL:

    • driver.documentURI(), document.documentURI:

    • driver.baseURI(), document.baseURI:

    • document.location:

    In summary, document.URL and document.documentURI are read-only properties that reflect the current URL of the document, while document.location is a read-write property that not only reflects the current URL but also allows you to navigate to a new one. The URL displayed in the address bar is a user-facing representation of the current document's URL, which is usually in sync with document.location.

    In addition to the above properties, The method driver.referrer() returns the document's referrer. The document.referrer property returns the URI of the page that linked to the current page. If the user navigated directly to the page (e.g., via a bookmark), the value is an empty string. Inside an <iframe>, the referrer is initially set to the same value as the href of the parent window's Window.location.

    The following example demonstrates how to use the WebDriver interface to visit a webpage and interact with it:

    fun visit() {
        val url = "https://twitter.com/home"
        val args = "-refresh"
        val options = session.options(args)
    
        options.eventHandlers.browseEventHandlers.onDocumentSteady.addLast { page, driver ->
            interact(page, driver)
        }
    
        session.load(url, options)
    }
    
    // interact with the page
    private suspend fun interact(driver: WebDriver) {
        val selector = "input[placeholder*=搜索], input[placeholder*=Search]"
        driver.waitForSelector(selector)
        driver.fill(selector, "Facebook")
        driver.press(selector, "Space")
        "Email".forEach { driver.press(selector, "$it") }
        driver.press(selector, "Enter")
    }
    • Constructor Detail

    • Method Detail

      • addInitScript

         abstract Unit addInitScript(String script)

        Adds a script which would be evaluated whenever the page is navigated. @mcp

        The script is evaluated after the document was created but before any of its scripts were run. This is useful to amend the JavaScript environment, e.g. to seed Math.random.

        This function should be invoked before a navigation, usually in an onWillNavigate event handler.

        Parameters:
        script - JavaScript source code to add.
      • addBlockedURLs

         abstract Unit addBlockedURLs(List<String> urlPatterns)

        Blocks resource URLs from loading. @mcp

        Parameters:
        urlPatterns - URL patterns to block.
      • open

         Unit open(String url)

        Opens the specified URL in the web driver. @mcp

        This function navigates the web driver to the provided URL and waits for the navigation to complete. It is a suspend function, meaning it can be used within coroutines for asynchronous execution.

        Example usage:

        driver.open("https://www.example.com")
        Parameters:
        url - The URL to which the web driver should navigate.
      • navigate

         abstract Unit navigate(String url)

        Navigates current page to the given URL. @mcp

        driver.navigate("https://www.example.com")
        driver.waitForNavigation()
        Parameters:
        url - URL to navigate page to.
      • navigate

         abstract Unit navigate(NavigateEntry entry)

        Navigates current page to the given URL. @mcp

        val entry = NavigateEntry("https://www.example.com?timestamp=11712067353", pageUrl = "https://www.example.com")
        driver.navigate(entry)
        driver.waitForNavigation()
        Parameters:
        entry - NavigateEntry to navigate page to.
      • reload

         abstract Unit reload()

        Reloads the current page. @mcp

        This method should trigger a reload of the current page, similar to how a user would refresh the page in a browser.

      • goBack

         abstract Unit goBack()

        Navigates the browser to the previous page in the navigation history. @mcp

        This method is expected to use the browser's navigation history to move back to the previous page. It should handle any exceptions that may occur during the navigation process.

      • goForward

         abstract Unit goForward()

        Navigates the browser to the next page in the navigation history. @mcp

        This method is expected to use the browser's navigation history to move forward to the next page. It should handle any exceptions that may occur during the navigation process.

      • currentUrl

         abstract String currentUrl()

        Returns a string representing the current URL that the browser is looking at. @mcp

        The current url is always the main frame's document.documentURI if the browser succeed to return it, and is displayed in the browser's address bar.

        If the browser failed to return a proper url, returns the passed in url to navigate, just like a real user enter a url in the address bar but the browser failed to load the page.

        NOTE: the url can be non-standard, for example:

        about:blank chrome://newtab chrome://settings

        see ai.platon.cdt.kt.protocol.types.page.NavigationEntry.userTypedURL

        Returns:

        A string containing the URL of the document, or the passed in url to navigate.

      • url

         abstract String url()

        The URL read-only property of the Document interface returns the document location as a string. @mcp

        This property equals to javascript document.URL. The document.URL property returns the same value. The document.documentURI property can be used on any document types, while the document.URL property can only be used on HTML documents.

        Returns:

        A string containing the URL of the document

      • documentURI

         abstract String documentURI()

        Returns the document location as a string. @mcp

        This property equals to javascript document.documentURI.

        The document.URL property which returns the same value. The document.documentURI property can be used on any document types, while the document.URL property can only be used on HTML documents.

        Returns:

        The document's documentURI.

      • baseURI

         abstract String baseURI()

        Returns the document's baseURI. @mcp

        The baseURI is a property of Node, it's the absolute base URL of the document containing the node. A baseURI is used to resolve relative URLs.

        The base URL is determined as follows:

        • By default, the base URL is the location of the document (as determined by window.location).

        • If the document has an <base> element, its href attribute is used.

        Returns:

        The document's baseURI.

      • referrer

         abstract String referrer()

        The referrer property returns the URI of the page that linked to this page. @mcp

        The value is an empty string if the user navigated to the page directly (not through a link, but, for example, by using a bookmark).

        Inside an <iframe>, the referrer will initially be set to the same value as the href of the parent window's Window.location.

        Returns:

        The document's referrer.

      • pageSource

         abstract String pageSource()

        Returns the source of the last loaded page. @mcp

        If the page has been modified after loading (for example, by JavaScript) there is no guarantee that the returned text is that of the modified page.

        PageSource and outerHTML:

        • pageSource: returns document HTML markup, will support non-HTML document

        • outerHTML: returns document HTML markup, for HTML document only

        val pageSource = driver.pageSource()
        Returns:

        The source of the current page

      • title

         abstract String title()

        Returns the title of the current page. @mcp

      • nanoDOMTree

         abstract <Error class: unknown class> nanoDOMTree()

        Retrieves a lightweight version of the DOM tree (NanoDOMTree). @mcp

        The NanoDOMTree is based on the accessibility tree and enhanced with DOM and document snapshot information. It is designed to be a compact representation suitable for AI processing and analysis.

        Returns:

        A NanoDOMTree representing the current page state, or null if retrieval fails.

      • chat

         abstract ModelResponse chat(String prompt, String selector)

        Interact with an AI model using the context of the element selected by selector. @mcp

        The AI model receives the element's content and structure as context along with the provided prompt.

        Parameters:
        prompt - The question or instruction for the AI model.
        selector - The CSS selector of the element to use as context.
        Returns:

        A ModelResponse containing the model's answer.

      • getCookies

         abstract List<Map<String, String>> getCookies()

        Returns the cookies of the current page. @mcp

        val cookies = driver.getCookies()
        Returns:

        The cookies of the current page.

      • deleteCookies

         abstract Unit deleteCookies(String name, String url, String domain, String path)

        Deletes browser cookies with matching name and url or domain/path pair. @mcp

        driver.deleteCookies("name", "https://www.example.com")

        NOTE: At least one of the url and domain needs to be specified

        Parameters:
        name - Name of the cookies to remove.
        url - If specified, deletes all the cookies with the given name where domain and path match provided URL.
        domain - If specified, deletes only cookies with the exact domain.
        path - If specified, deletes only cookies with the exact path.
      • waitForSelector

         Duration waitForSelector(String selector)

        Wait until the element identified by the selector becomes present in the DOM or timeout. @mcp

        val remainingTime = driver.waitForSelector("h2.title")
        Parameters:
        selector - The selector of the element to wait for.
        Returns:

        The remaining time until timeout when the element becomes present.

      • waitForSelector

         Long waitForSelector(String selector, Long timeoutMillis)

        Wait until the element identified by the selector becomes present in the DOM or timeout. @mcp

        val remainingTime = driver.waitForSelector("h2.title", 30000)
        Parameters:
        timeoutMillis - The maximum time to wait for the element to become present.
        Returns:

        The remaining time until timeout when the element becomes present.

      • waitForSelector

         Duration waitForSelector(String selector, Duration timeout)

        Wait for the element identified by the selector to become present in the DOM, or until timeout. @mcp

        val remainingTime = driver.waitForSelector("h2.title", Duration.ofSeconds(30))
        Parameters:
        timeout - The maximum time to wait for the element to become present.
        Returns:

        The remaining time until timeout when the element becomes present.

      • waitForSelector

         abstract Duration waitForSelector(String selector, SuspendFunction0<Unit> action)

        Wait for the element identified by the selector to become present in the DOM, or until timeout. @mcp

        This method periodically checks for the existence of the element. If the element is not found during a check, the action will be executed, such as scrolling the page down.

        val remainingTime = driver.waitForSelector("h2.title") {
         driver.scrollDown()
        }
        Parameters:
        action - The action to execute when the element is not found.
      • waitForSelector

         Long waitForSelector(String selector, Long timeoutMillis, SuspendFunction0<Unit> action)

        Wait for the element identified by the selector to become present in the DOM, or until timeout. @mcp

        This method periodically checks for the existence of the element. If the element is not found during a check, the action will be executed, such as scrolling the page down.

        val remainingTime = driver.waitForSelector("h2.title", 30000) {
         driver.scrollDown()
        }
        Parameters:
        timeoutMillis - The maximum time to wait for the element to become present.
        action - The action to execute when the element is not found.
        Returns:

        The remaining time until timeout when the element becomes present.

      • waitForSelector

         abstract Duration waitForSelector(String selector, Duration timeout, SuspendFunction0<Unit> action)

        Wait for the element identified by the selector to become present in the DOM, or until timeout. @mcp

        This method periodically checks for the existence of the element. If the element is not found during a check, the action will be executed, such as scrolling the page down.

        val remainingTime = driver.waitForSelector("h2.title", Duration.ofSeconds(30)) {
         driver.scrollDown()
        }
        Parameters:
        timeout - The maximum time to wait for the element to become present.
        action - The action to execute when the element is not found.
        Returns:

        The remaining time until timeout when the element becomes present.

      • waitForNavigation

         abstract Duration waitForNavigation(String oldUrl)

        Wait until the current url changes or timeout. @mcp

        val url = "https://www.example.com"
        driver.navigate(url)
        var remainingTime = driver.waitForNavigation()
        if (remainingTime 0) {
          driver.click("a[href='/next']")
          remainingTime = driver.waitForNavigation(url)
        }
      • waitForNavigation

         Long waitForNavigation(String oldUrl, Long timeoutMillis)

        Wait until the current url changes or timeout. @mcp

        val url = "https://www.example.com"
        driver.navigate(url)
        var remainingTime = driver.waitForNavigation(1000)
        if (remainingTime 0) {
          driver.click("a[href='/next']")
          remainingTime = driver.waitForNavigation(url, 1000)
        }
        Parameters:
        timeoutMillis - The maximum time to wait for the url to change.
      • waitForNavigation

         abstract Duration waitForNavigation(String oldUrl, Duration timeout)

        Wait until the current url changes or timeout. @mcp

        val timeout = Duration.ofSeconds(30)
        val url = "https://www.example.com"
        driver.navigate(url)
        var remainingTime = driver.waitForNavigation(timeout)
        if (remainingTime 0) {
          driver.click("a[href='/next']")
          remainingTime = driver.waitForNavigation(url, timeout)
        }
        Parameters:
        timeout - The maximum time to wait for the url to change.
      • waitForPage

         abstract WebDriver waitForPage(String url, Duration timeout)

        Await navigation to the specified URL page or timeout if necessary. @mcp

        val newDriver = driver.waitForPage("https://www.example.com", Duration.ofSeconds(30))

        TODO: check if waitForPage and waitForNavigation can be merged into one method.

        Parameters:
        url - The URL to navigate to.
        Returns:

        The remaining time until timeout when the predicate returns true.

      • waitForFunction

         abstract WebDriver waitForFunction(String pageFunction, Duration timeout)

        Returns when the pageFunction returns a truthy value. @mcp

        Parameters:
        pageFunction - A JavaScript function to be evaluated in the page context.
      • waitUntil

         abstract Duration waitUntil(SuspendFunction0<Boolean> predicate)

        Wait until the predicate returns true. @mcp

        val remainingTime = driver.waitUntil {
           driver.exists("h2.title")
        }
        Parameters:
        predicate - The predicate to check.
        Returns:

        The remaining time until timeout when the predicate returns true.

      • waitUntil

         Long waitUntil(Long timeoutMillis, SuspendFunction0<Boolean> predicate)

        Wait until the predicate returns true. @mcp

        val remainingTime = driver.waitUntil(10000) {
          driver.exists("h2.title")
        }
        Parameters:
        timeoutMillis - The maximum time to wait for the predicate to return true.
        predicate - The predicate to check.
        Returns:

        The remaining time until timeout when the predicate returns true.

      • waitUntil

         abstract Duration waitUntil(Duration timeout, SuspendFunction0<Boolean> predicate)

        Wait until the predicate returns true. @mcp

        val remainingTime = driver.waitUntil(Duration.ofSeconds(10)) {
           driver.exists("h2.title")
        }
        Parameters:
        timeout - The maximum time to wait for the predicate to return true.
        predicate - The predicate to check.
        Returns:

        The remaining time until timeout when the predicate returns true.

      • exists

         abstract Boolean exists(String selector)

        Returns whether the element exists. @mcp

        driver.exists("h2.title")
        Parameters:
        selector -
        • The selector of the element to check.

        Returns:

        Whether the element exists.

      • isHidden

         Boolean isHidden(String selector)

        Returns whether the element is hidden. @mcp

        driver.isHidden("input[name='q']")
        Parameters:
        selector -
        • The selector of the element to check.

        Returns:

        Whether the element is hidden.

      • isVisible

         abstract Boolean isVisible(String selector)

        Returns whether the element is visible. @mcp

        driver.isVisible("input[name='q']")
        Parameters:
        selector -
        • The selector of the element to check.

        Returns:

        Whether the element is visible.

      • isChecked

         abstract Boolean isChecked(String selector)

        Returns whether the element is checked. @mcp

        driver.isChecked("input[name='agree']")
        Parameters:
        selector -
        • The selector of the element to check.

        Returns:

        Whether the element is checked.

      • bringToFront

         abstract Unit bringToFront()

        Brings the browser window to the front. @mcp

        driver.bringToFront()
      • hover

         abstract Unit hover(String selector)

        This method hovers over the element. @mcp

        • Scroll the element into view if needed.

        • Move mouse to hover over a random position near the center of the element.

        driver.hover("h1")
      • focus

         abstract Unit focus(String selector)

        This method fetches an element with selector and focuses it. @mcp

        If there's no element matching selector, nothing to do.

        driver.focus("input[name='q']")
        Parameters:
        selector -
        • A selector of an element to focus. If there are multiple elements satisfying the selector, the first will be focused.

      • type

         abstract Unit type(String selector, String text)

        This method emulates inserting text that doesn't come from a key press. @mcp

        driver.type("input[name='q']", "Hello, World!")
        Parameters:
        selector -
        • A selector of an element to focus. If there are multiple elements satisfying the selector, the first will be focused.

        text - The text to insert.
      • fill

         abstract Unit fill(String selector, String text)

        This method emulates inserting text that doesn't come from a key press. @mcp

        Unlike type, this method clears the existing value before typing.

        driver.fill("input[name='q']", "Hello, World!")
        Parameters:
        selector -
        • A selector of an element to focus, and then fill text into it. If there are multiple elements satisfying the selector, the first will be focused.

        text - The text to fill.
      • press

         abstract Unit press(String selector, String key)

        Shortcut for keyboard down and keyboard up. @mcp

        The key is specified as a string, which can be a single character, a key name, or a combination of both. For example, 'a', 'A', 'KeyA', 'Enter', 'Shift+A', and 'Control+Shift+Tab' are all valid keys.

        driver.press("input[name='q']", "Enter")
        Parameters:
        selector -
        • A selector of an element to focus, and then press a key. If there are multiple elements satisfying the selector, the first will be focused.

        key -
      • keyDown

         abstract Unit keyDown(String key)

        Presses and holds a keyboard key on the currently focused element. @mcp

        driver.keyDown("Shift")
        driver.keyDown("Control")
        Parameters:
        key - A key to press and hold.
      • keyUp

         abstract Unit keyUp(String key)

        Releases a previously pressed keyboard key on the currently focused element. @mcp

        driver.keyUp("Shift")
        driver.keyUp("Control")
        Parameters:
        key - A key to release.
      • click

         abstract Unit click(String selector, Integer count)

        Focus on an element with selector and click it. @mcp

        If there's no element matching selector, nothing to do.

        driver.click("button[type='submit']")
        Parameters:
        selector -
        • A selector of an element to focus. If there are multiple elements satisfying the selector, the first will be focused.

        count - The number of times to click.
      • click

         abstract Unit click(String selector, String modifier)

        Focus on an element with selector and click it with modifier pressed. @mcp

        Parameters:
        selector - The selector of the element to click.
        modifier - The keyboard modifier to press while clicking (e.g., "Shift", "Control", "Alt").
      • dblclick

         abstract Unit dblclick(String selector)

        Focus on an element with selector and double-click it. @mcp

        Parameters:
        selector - The selector of the element to double-click.
      • dblclick

         abstract Unit dblclick(String selector, String modifier)

        Focus on an element with selector and double-click it with modifier pressed. @mcp

        Parameters:
        selector - The selector of the element to double-click.
        modifier - The keyboard modifier to press while double-clicking.
      • resize

         abstract Unit resize(Integer width, Integer height)

        Resizes the viewport to the specified width and height. @mcp

      • clickTextMatches

         abstract Unit clickTextMatches(String selector, String pattern, Integer count)

        This method clicks an element with selector whose text content matches pattern, and then focuses it. @mcp

        If there's no element matching selector, or the element's text content doesn't match pattern, nothing to do.

        driver.clickTextMatches("button", "submit")

        TODO: use playwright style selector which supports text matching, for example: button:has-text("submit")

        Parameters:
        selector -
        • A selector of an element to focus. If there are multiple elements satisfying the selector, the first will be focused.

        pattern - The pattern to match the text content.
        count - The number of times to click.
      • clickMatches

         abstract Unit clickMatches(String selector, String attrName, String pattern, Integer count)

        This method clicks an element with selector whose attribute name is attrName and value matches pattern, and then focuses it. @mcp

        If there's no element matching selector, or the element has no attribute attrName, or the element's attribute value doesn't match pattern, nothing to do.

        driver.clickMatches("button", "type", "submit")
        Parameters:
        selector -
        • A selector of an element to focus. If there are multiple elements satisfying the selector, the first will be focused.

        attrName - The attribute name to match.
        pattern - The pattern to match the attribute value.
        count - The number of times to click.
      • selectOption

         abstract List<String> selectOption(String selector, List<String> values)

        Selects one or more options in a <select> element. @mcp

        Parameters:
        selector - A selector to query the element
        values - The values or labels of the options to select
        Returns:

        The list of selected option values

      • selectOption

         String selectOption(String selector, String value)

        Selects an option in a <select> element. @mcp

        Parameters:
        selector - A selector to query the element
        value - The value or label of the option to select
        Returns:

        The selected option value, or null if not found

      • check

         abstract Unit check(String selector)

        This method check an element with selector. If there's no element matching selector, nothing to do. @mcp

        Parameters:
        selector -
        • A selector of an element to check. If there are multiple elements satisfying the selector, the first will be checked.

      • uncheck

         abstract Unit uncheck(String selector)

        This method uncheck an element with selector. If there's no element matching selector, nothing to do. @mcp

        Parameters:
        selector -
        • A selector of an element to uncheck. If there are multiple elements satisfying the selector, the first will be focused.

      • scrollTo

         abstract Double scrollTo(String selector)

        This method fetches an element with selector, scrolls it into view if needed. @mcp

        If there's no element matching selector, the method does nothing.

        driver.scrollTo("h2.title")
        Parameters:
        selector -
        • A selector to search for element to scroll to. If there are multiple elements satisfying the selector, the first will be selected.

      • scrollDown

         abstract Double scrollDown(Integer count)

        The current page frame scrolls down for count times. @mcp

        driver.scrollDown(3)

        TODO: use mouseWheel instead

        Parameters:
        count - The times to scroll down.
      • scrollUp

         abstract Double scrollUp(Integer count)

        The current page frame scrolls up for count times. @mcp

        driver.scrollUp(3)

        TODO: use mouseWheel instead

        Parameters:
        count - The times to scroll up.
      • scrollBy

         abstract Double scrollBy(Double pixels, Boolean smooth)

        Scrolls the current page frame vertically by the given number of pixels. @mcp

        Positive values scroll down, negative values scroll up. The target position is clamped to the scrollable range 0, scrollHeight - viewportHeight.

        When smooth is true a deterministic, step‑wise scrolling algorithm is used (instead of relying on the browser's native smooth scrolling) for higher stability: it issues incremental window.scrollTo() calls and waits for the viewport to settle after each step. This reduces race conditions with lazy‑loaded content or async layout changes and yields more predictable final positions.

        Typical usage:

        // Scroll down 200px smoothly (default)
        driver.scrollBy()
        // Scroll up 500px immediately
        driver.scrollBy(-500.0, smooth = false)
        // Scroll down one viewport height, smooth
        val y = driver.scrollBy(driver.viewportHeight(), smooth = true)
        Parameters:
        pixels - The delta in pixels to scroll; positive scrolls down, negative scrolls up.
        smooth - Whether to perform a deterministic smooth scrolling sequence.
        Returns:

        The final vertical scroll offset (window.scrollY) after scrolling.

      • scrollToTop

         abstract Double scrollToTop()

        The current page frame scrolls to the top. @mcp

        driver.scrollToTop()
      • scrollToBottom

         abstract Double scrollToBottom()

        The current page frame scrolls to the bottom. @mcp

        driver.scrollToBottom()
      • scrollToMiddle

         abstract Double scrollToMiddle(Double ratio)

        The current page frame scrolls to the middle. @mcp

        driver.scrollToMiddle(0.2)
        driver.scrollToMiddle(0.5)
        driver.scrollToMiddle(0.8)
        Parameters:
        ratio - The ratio of the page to scroll to, 0.0 means the top, 1.0 means the bottom.
      • scrollToViewport

         abstract Double scrollToViewport(Double n, Boolean smooth)

        Scroll to the 2.5th viewport position. @mcp

        driver.scrollToViewport(1.0)
        driver.scrollToViewport(1.5)
        driver.scrollToViewport(2.5)
        driver.scrollToViewport(3.0)
        Parameters:
        n - The viewport number of the page to scroll to (1-based).
      • mouseWheelDown

        @Deprecated(message = "Use mouseWheel(deltaX, deltaY) instead", replaceWith = @ReplaceWith(imports = {}, expression = "mouseWheel(deltaX, deltaY)")) abstract Unit mouseWheelDown(Integer count, Double deltaX, Double deltaY, Long delayMillis)

        The mouse wheels down for count times. @mcp

        driver.mouseWheelDown(3)
        Parameters:
        count - The times to wheel down.
        deltaX - The distance to wheel horizontally.
        deltaY - The distance to wheel vertically.
        delayMillis - The delay time in milliseconds.
      • mouseWheelUp

        @Deprecated(message = "Use mouseWheel(deltaX, deltaY) instead", replaceWith = @ReplaceWith(imports = {}, expression = "mouseWheel(deltaX, deltaY)")) abstract Unit mouseWheelUp(Integer count, Double deltaX, Double deltaY, Long delayMillis)

        The mouse wheels up for count times. @mcp

        driver.mouseWheelUp(3)
        Parameters:
        count - The times to wheel up.
        deltaX - The distance to wheel horizontally.
        deltaY - The distance to wheel vertically.
        delayMillis - The delay time in milliseconds.
      • mouseWheel

         abstract Unit mouseWheel(Double deltaX, Double deltaY)

        Scrolls the mouse wheel by the provided deltas. @mcp

        Positive deltaY scrolls down and negative deltaY scrolls up.

        driver.mouseWheel(0.0, 150.0)
        driver.mouseWheel(0.0, -150.0)
        Parameters:
        deltaX - The distance to wheel horizontally.
        deltaY - The distance to wheel vertically.
      • moveMouseTo

        @Deprecated(message = "Use mouseMove instead", replaceWith = @ReplaceWith(imports = {}, expression = "mouseMove(x, y)")) abstract Unit moveMouseTo(Double x, Double y)

        The mouse moves to the position specified by x and y. @mcp

        driver.moveMouseTo(100.0, 200.0)
        Parameters:
        x - The x coordinate to move to.
        y - The y coordinate to move to.
      • moveMouseTo

         abstract Unit moveMouseTo(String selector, Integer deltaX, Integer deltaY)

        The mouse moves to the element with selector. @mcp

        driver.moveMouseTo("h2.title")
        Parameters:
        deltaX - The distance to the left of the element.
        deltaY - The distance to the top of the element.
      • mouseMove

         abstract Unit mouseMove(Double x, Double y)

        Moves the mouse to the position specified by x and y. @mcp

        driver.mouseMove(100.0, 200.0)
        Parameters:
        x - The x coordinate to move to.
        y - The y coordinate to move to.
      • mouseDown

         abstract Unit mouseDown(String button, Integer clickCount)

        Presses a mouse button at the current mouse position. @mcp

        driver.mouseDown()
        driver.mouseDown("right")
        Parameters:
        button - The mouse button to press.
        clickCount - The click count associated with the press event.
      • mouseUp

         abstract Unit mouseUp(String button, Integer clickCount)

        Releases a mouse button at the current mouse position. @mcp

        driver.mouseUp()
        driver.mouseUp("right")
        Parameters:
        button - The mouse button to release.
        clickCount - The click count associated with the release event.
      • dragAndDrop

         abstract Unit dragAndDrop(String selector, Integer deltaX, Integer deltaY)

        Performs a drag, dragenter, dragover, and drop in sequence. @mcp

        Parameters:
        selector -
        • selector of the element to drag from.

        deltaX - The distance to drag horizontally.
        deltaY - The distance to drag vertically.
      • drag

         Unit drag(String sourceSelector, String targetSelector)

        Drags the element identified by sourceSelector onto the element identified by targetSelector. @mcp

        The implementation dispatches an HTML5 drag sequence between the source and target elements using a shared DataTransfer payload so selector-to-selector drag flows can be asserted reliably by automation clients.

        Parameters:
        sourceSelector - The selector of the element to drag.
        targetSelector - The selector of the element to drop onto.
      • outerHTML

         abstract String outerHTML()

        Returns the document's HTML markup. @mcp

        val html = driver.outerHTML()

        If the document does not exist, returns null.

        Returns:

        The HTML markup of the document.

      • outerHTML

         abstract String outerHTML(String selector)

        Returns the node's HTML markup, the node is located by selector. @mcp

        If the node does not exist, returns null.

        val html = driver.outerHTML("h2.title")
        Parameters:
        selector - The selector to locate the node.
        Returns:

        The HTML markup of the node.

      • textContent

         abstract String textContent(String selector)

        Returns the text content of the document or a specific element. @mcp

        If selector is null, returns the text content of the entire document (usually document.body.innerText). If selector is provided, returns the text content of the first matching element.

        If the document or element does not exist, returns null.

        val bodyText = driver.textContent()
        val titleText = driver.textContent("h1.title")
        Parameters:
        selector - The CSS selector of the element.
        Returns:

        The text content, or null if not found.

      • extract

         abstract Map<String, String> extract(Map<String, String> fields)

        Extracts multiple fields from the page using CSS selectors. @mcp

        Parameters:
        fields - A map where keys are field names and values are CSS selectors.
        Returns:

        A map containing the extracted text content for each field. Values are null if the selector matches nothing.

      • querySelectorAll

        @Beta() abstract List<NodeRef> querySelectorAll(String selector)

        Use querySelectorAll to get all matched elements, and then return their NodeRefs. @mcp

        Note that the NodeRefs may become stale after certain operations, so they should be used immediately after selection.

        Parameters:
        selector - selector string, supported selector types include CSS selector, XPath, and backend node id.
        Returns:

        a list of NodeRefs for the matched elements, or an empty list if no elements are matched or an error occurs.

      • selectFirstTextOrNull

         abstract String selectFirstTextOrNull(String selector)

        Returns the node's text content, the node is located by selector. @mcp

        If the node does not exist, returns null.

        val text = driver.selectFirstTextOrNull("h2.title")
        Parameters:
        selector - The selector to locate the node.
        Returns:

        The text content of the node.

      • selectTextAll

         abstract List<String> selectTextAll(String selector)

        Returns a list of text contents of all the elements matching the specified selector within the page. @mcp

        If no elements match the selector, returns an empty list.

        val texts = driver.selectTextAll("h2")
        Parameters:
        selector - The selector to locate the nodes.
        Returns:

        The text contents of the nodes.

      • selectFirstAttributeOrNull

         abstract String selectFirstAttributeOrNull(String selector, String attrName)

        Returns the node's attribute value, the node is located by selector, the attribute is attrName. @mcp

        If the node does not exist, or the attribute does not exist, returns null.

        val classes = driver.selectFirstAttributeOrNull("h2.title", "class")
        Parameters:
        selector - The selector to locate the node.
        attrName - The attribute name to retrieve.
        Returns:

        The attribute value of the node.

      • selectAttributes

         abstract Map<String, String> selectAttributes(String selector)

        Returns the node's attribute values, the node is located by selector. @mcp

        If the node do not exist, or the attribute does not exist, returns an empty list.

        val classes = driver.selectAttributes("h2.title", "class")
        Parameters:
        selector - The selector to locate the nodes.
        Returns:

        The attribute pairs of the nodes.

      • selectAttributeAll

         abstract List<String> selectAttributeAll(String selector, String attrName, Integer start, Integer limit)

        Returns the nodes' attribute values, the nodes are located by selector, the attribute is attrName. @mcp

        If the nodes do not exist, or the attribute does not exist, returns an empty list.

        val classes = driver.selectAttributeAll("h2.title", "class")
        Parameters:
        selector - The selector to locate the nodes.
        attrName - The attribute name to retrieve.
        start - The offset of the first node to select.
        limit - The maximum number of nodes to select.
        Returns:

        The attribute values of the nodes.

      • setAttribute

         abstract Unit setAttribute(String selector, String attrName, String attrValue)

        Set the attribute of an element located by selector. @mcp

        driver.setAttribute("h2.title", "class", "header")
        Parameters:
        selector - The CSS query to select an element.
        attrName - The attribute name to set.
        attrValue - The attribute value to set.
      • setAttributeAll

         abstract Unit setAttributeAll(String selector, String attrName, String attrValue)

        Set the attribute of all elements matching the CSS query. @mcp

        driver.setAttributeAll("h2.title", "class", "header")
        Parameters:
        selector - The CSS query to select elements.
        attrName - The attribute name to set.
        attrValue - The attribute value to set.
      • selectFirstPropertyValueOrNull

         abstract String selectFirstPropertyValueOrNull(String selector, String propName)

        Returns the node's property value, the node is located by selector, the property is propName. @mcp

        If the node does not exist, or the property does not exist, returns null.

        val classes = driver.selectFirstPropertyOrNull("input#input", "value")
        Parameters:
        selector - The selector to locate the node.
        propName - The property name to retrieve.
        Returns:

        The property value of the node.

      • selectPropertyValueAll

         abstract List<String> selectPropertyValueAll(String selector, String propName, Integer start, Integer limit)

        Returns the nodes' property values, the nodes are located by selector, the property is propName. @mcp

        If the nodes do not exist, or the property does not exist, returns an empty list.

        val classes = driver.selectPropertyAll("input#input", "value")
        Parameters:
        selector - The selector to locate the nodes.
        propName - The property name to retrieve.
        start - The offset of the first node to select.
        limit - The maximum number of nodes to select.
        Returns:

        The property values of the nodes.

      • setProperty

         abstract Unit setProperty(String selector, String propName, String propValue)

        Set the property of an element located by selector. @mcp

        driver.setProperty("input#input", "value")
        Parameters:
        selector - The CSS query to select an element.
        propName - The property name to set.
        propValue - The property value to set.
      • setPropertyAll

         abstract Unit setPropertyAll(String selector, String propName, String propValue)

        Set the property of all elements matching the CSS query. @mcp

        driver.setPropertyAll("input#input", "value")
        Parameters:
        selector - The CSS query to select elements.
        propName - The property name to set.
        propValue - The property value to set.
      • selectHyperlinks

         abstract List<Hyperlink> selectHyperlinks(String selector, Integer offset, Integer limit)

        Find hyperlinks in elements matching the CSS query.

        val hyperlinks = driver.selectHyperlinks("a.product-link")
        Parameters:
        selector - The CSS query to select elements.
        offset - The offset of the first element to select.
        limit - The maximum number of elements to select.
        Returns:

        The hyperlinks in the elements.

      • selectAnchors

         abstract List<GeoAnchor> selectAnchors(String selector, Integer offset, Integer limit)

        Find anchor elements matching the CSS query.

        val anchors = driver.selectAnchors("a.product-link")
        Parameters:
        selector - The CSS query to select elements.
        offset - The offset of the first element to select.
        limit - The maximum number of elements to select.
        Returns:

        The anchors.

      • selectImages

         abstract List<String> selectImages(String selector, Integer offset, Integer limit)

        Find image elements matching the CSS query.

        val images = driver.selectImages("img.product-image")
        Parameters:
        selector - The CSS query to select elements.
        offset - The offset of the first element to select.
        limit - The maximum number of elements to select.
        Returns:

        The image URLs.

      • evaluate

         abstract Object evaluate(String expression)

        Executes JavaScript in the context of the currently selected frame or window. @mcp

        If the result is not JavaScript object, it is not returned.

        If you want to execute a function, convert it to IIFE (Immediately Invoked Function Expression).

        val title = driver.evaluate("document.title")

        Multi-line JavaScript code:

        val code = """
        () => {
          const a = 10;
          const b = 20;
          return a * b;
        }
        """.trimIndent()
        
        val result = driver.evaluate(code)
        • Wrap the code in an IIFE (Immediately Invoked Function Expression) to return a value.

        • Escape line breaks with \n.

        Parameters:
        expression - JavaScript expression to evaluate
        Returns:

        Remote object value in case of primitive values or null.

      • evaluate

         abstract <T extends Any> T evaluate(String expression, T defaultValue)

        Executes JavaScript and returns the result, or defaultValue if the result is null or incompatible. @mcp

        Parameters:
        expression - The JavaScript expression to evaluate.
        defaultValue - The value to return if evaluation fails or returns null.
        Returns:

        The evaluation result or defaultValue.

      • evaluateDetail

         abstract JsEvaluation evaluateDetail(String expression)

        returns detailed evaluation metadata (beta). @mcp

        Parameters:
        expression - The JavaScript expression to evaluate.
        Returns:

        A JsEvaluation object containing the result and metadata.

      • evaluateValue

         abstract Object evaluateValue(String expression)

        Executes JavaScript and returns the result as a JSON-serializable value. @mcp

        If the result is an object, it is serialized to a Map or List. If the result is a primitive, it is returned as is. If the result is null or undefined, null is returned.

        Parameters:
        expression - The JavaScript expression to evaluate.
        Returns:

        The result as a JSON-compatible object (Map, List, String, Number, Boolean, or null).

      • evaluateValue

         abstract <T extends Any> T evaluateValue(String expression, T defaultValue)

        Executes JavaScript and returns the result as a JSON-serializable value, or defaultValue if null/incompatible. @mcp

        Parameters:
        expression - The JavaScript expression to evaluate.
        defaultValue - The value to return if evaluation fails or returns null.
        Returns:

        The evaluation result or defaultValue.

      • evaluateValueDetail

         abstract JsEvaluation evaluateValueDetail(String expression)

        Returns detailed value evaluation metadata (beta). @mcp

        Parameters:
        expression - The JavaScript expression to evaluate.
        Returns:

        A JsEvaluation object containing the result value and metadata.

      • screenshot

         abstract String screenshot(Boolean fullPage)

        Capture a screenshot of the current viewport (or primary browsing context) after ensuring any pending layout. @mcp

        If the backend supports element-centric capture this may represent the full page; implementation specific.

        The target element (if any) is scrolled into view before capture.

        val base64 = driver.screenshot()
        val bytes = Base64.getDecoder().decode(base64)
      • screenshot

         abstract String screenshot(String selector)

        Scroll the element matched by selector into view (if needed) then take a screenshot of that element's bounding box. @mcp

        Returns a Base64 encoded image (implementation usually PNG/JPEG), PNG by default.

      • screenshot

         abstract String screenshot(RectD rect)

        Take a screenshot of the rectangle specified by rect in the current page coordinate space. @mcp

        Caller is responsible for ensuring the rectangle is visible or scrolled into view if the implementation requires it.

      • ariaSnapshot

         String ariaSnapshot(String viewports)

        Return the ARIA snapshot (accessibility tree in YAML format) for the specified viewports. @mcp

        The viewports parameter controls which viewport(s) to include:

        • "all" — return the full-page snapshot (default).

        • "3" — return only viewport 3.

        • "1,3,5" — return viewports 1, 3 and 5.

        • "2-4" — return viewports 2, 3 and 4.

        Viewport indices are 1-based. The viewport height is determined by the browser's current viewport dimensions.

        Parameters:
        viewports - A viewport specification string (e.g.
        Returns:

        The ARIA snapshot YAML covering only the requested viewports.

      • clickablePoint

         abstract PointD clickablePoint(String selector)

        Calculate the clickable point of an element located by selector. @mcp

        If the element does not exist, or is not clickable, returns null.

        Parameters:
        selector - The selector of the element to calculate the clickable point.
        Returns:

        The clickable point of the element.

      • boundingBox

         abstract RectD boundingBox(String selector)

        Return the bounding box of an element located by selector. @mcp

        If the element does not exist, returns null.

        Parameters:
        selector - The selector of the element to calculate the bounding box.
        Returns:

        The bounding box of the element.

      • newJsoupSession

         abstract Connection newJsoupSession()

        Create a new Jsoup session with the last page's context, which means, the same headers and cookies. @mcp

        Returns:

        The Jsoup session.

      • loadJsoupResource

         abstract Connection.Response loadJsoupResource(String url)

        Load the url as a resource with Jsoup rather than browser rendering, with the last page's context. @mcp

        This means, the same headers and cookies.

        Parameters:
        url - The URL to load.
        Returns:

        The Jsoup response.

      • loadResource

         abstract NetworkResourceResponse loadResource(String url)

        Load the url as a resource without browser rendering, with the last page's context. @mcp

        This means, the same headers and cookies.

        Parameters:
        url - The URL to load.
        Returns:

        The network resource response.

      • delay

         <Error class: unknown class> delay(Long millis)

        Delay for a given amount of time. @mcp

        Parameters:
        millis - The amount of time to delay, in milliseconds.
      • delay

         <Error class: unknown class> delay(Duration duration)

        Delay for a given amount of time. @mcp

        Parameters:
        duration - The amount of time to delay.
      • delay

         <Error class: unknown class> delay(Duration duration)

        Delay for a given amount of time. @mcp

        Parameters:
        duration - The amount of time to delay.
      • upload

         abstract Unit upload(String selector, List<String> paths)

        Upload files to the element located by selector. @mcp

        Parameters:
        selector - The selector of the file input element.
        paths - The list of file paths to upload.
      • pause

         abstract Unit pause()

        Force the page pauses all navigations and PENDING resource fetches. @mcp

        If the page loading pauses, the user can still interact with the page, and therefore resources can continue to load.

      • stop

         abstract Unit stop()

        Force the page stop all navigations and RELEASES all resources. @mcp

        Interaction with the stop page results in undefined behavior and the results should not be trusted.

        If a web driver stops, it can later be used to visit other pages.

      • getId

         abstract Integer getId()

        The driver id, starts with 1. The id is unique in process scope.

      • getParentSid

         abstract Integer getParentSid()

        The parent driver id. The id is unique in process scope.

      • getBrowser

         abstract Browser getBrowser()

        The browser of the driver. The browser defines methods and events to manipulate a real browser.

      • getFrames

         abstract List<WebDriver> getFrames()

        Web pages for the page open from the current page, via window.open(), link click, form submission, etc.

        TODO: NOT IMPLEMENTED

      • getOutgoingPages

         abstract Set<WebDriver> getOutgoingPages()

        Web pages for the page open from the current page, via window.open(), link click, form submission, etc.

      • getBrowserType

         abstract BrowserType getBrowserType()

        The browser type. BrowserType.PULSAR_CHROME is the only fully supported browser type currently. BrowserType.PLAYWRIGHT_CHROME is a partially supported browser type since Playwright is not thread safe.

      • getDelayPolicy

         abstract Map<String, IntRange> getDelayPolicy()

        The delay policy of the driver. The delay policy is a map of delay ranges in milliseconds for different actions.

        The delay policy is used to simulate human behaviors, such as typing, clicking, scrolling, etc.

        delayPolicy["click"] == 500..1000
      • getTimeoutPolicy

         abstract Map<String, Duration> getTimeoutPolicy()

        The timeout policy of the driver. The timeout policy is a map of timeout durations for different actions.

        The timeout policy is used to set the maximum time to wait for an action to complete.

        timeoutPolicy["click"] == Duration.ofSeconds(30)