Appium Windows Driver is a test automation tool for Windows devices and acts as a proxy to Microsoft's WinAppDriver server. Appium Windows Driver supports testing Universal Windows Platform (UWP), Windows Forms (WinForms), Windows Presentation Foundation (WPF), and Classic Windows (Win32) apps on Windows 10 PCs. The server itself is maintained by Microsoft at https://github.com/microsoft/WinAppDriver. Check its release notes and the vendor documentation to get more details on the supported features and possible pitfalls.
Note
Since version 2.0.0 Windows driver has dropped the support of Appium 1, and is only compatible to Appium 2. Use the
appium driver install --source=npm appium-windows-driver
command to add it to your Appium 2 dist.
Note
This project is actively looking for maintainers with DotNet experience.
Beside of standard Appium requirements Appium Windows Driver adds the following prerequisites:
- Appium Windows Driver only supports Windows 10 as the host.
- Developer mode must be enabled
- Since version 3.0.0 this driver does not automatically install WinAppDriver server anymore. You should perform the server installation via the install-wad driver script instead. Driver versions below 3.0.0 download and install a bundled WinAppDriver package version automatically upon executing its installation via the Appium command line interface. Although, in such case the actual server binary version could be out of date. You could download and install the most recent version of WinAppDriver server manually from the GitHub releases page.
Appium Windows Driver supports the following capabilities:
Capability Name | Description |
---|---|
platformName | Must be set to windows (case-insensitive). |
appium:automationName | Must be set to windows (case-insensitive). |
appium:app | The name of the UWP application to test or full path to a classic app, for example Microsoft.WindowsCalculator_8wekyb3d8bbwe!App or C:\Windows\System32\notepad.exe . It is also possible to set app to Root . In such case the session will be invoked without any explicit target application (actually, it will be Explorer). Either this capability or appTopLevelWindow must be provided on session startup. |
appium:appArguments | Application arguments string, for example /? . |
appium:appTopLevelWindow | The hexadecimal handle of an existing application top level window to attach to, for example 0x12345 (should be of string type). Either this capability or app must be provided on session startup. |
appium:appWorkingDir | Full path to the folder, which is going to be set as the working dir for the application under test. This is only applicable for classic apps. |
appium:createSessionTimeout | Timeout in milliseconds used to retry Appium Windows Driver session startup. This capability could be used as a workaround for the long startup times of UWP applications (aka Failed to locate opened application window with appId: TestCompany.my_app4!App, and processId: 8480 ). Default value is 20000 . |
ms:waitForAppLaunch | Similar to createSessionTimeout , but in seconds and is applied on the server side. Enables Appium Windows Driver to wait for a defined amount of time after an app launch is initiated prior to attaching to the application session. The limit for this is 50 seconds. |
ms:experimental-webdriver | Enables experimental features and optimizations. See Appium Windows Driver release notes for more details on this capability. false by default. |
ms:forcequit | Defines if the WinAppDriver should be started with the /forcequit command line argument which will forcefully kill the application process during session termination. Default false . |
appium:systemPort | The port number to execute Appium Windows Driver server listener on, for example 5556 . The port must not be occupied. The default starting port number for a new Appium Windows Driver session is 4724 . If this port is already busy then the next free port will be automatically selected. |
appium:prerun | An object containing either script or command key. The value of each key must be a valid PowerShell script or command to be executed prior to the WinAppDriver session startup. See Power Shell commands execution for more details. Example: {script: 'Get-Process outlook -ErrorAction SilentlyContinue'} |
appium:postrun | An object containing either script or command key. The value of each key must be a valid PowerShell script or command to be executed after WinAppDriver session is stopped. See Power Shell commands execution for more details. |
appium:newCommandTimeout | How long (in seconds) the driver should wait for a new command from the client before assuming the client has stopped sending requests. After the timeout, the session is going to be deleted. 60 seconds by default. Setting it to zero disables the timer. |
This script is used to install the given or latest stable version of WinAppDriver server from
the GitHub releases page.
Run appium driver run windows install-wad <optional_wad_version>
, where optional_wad_version
must be either valid WAD version number or should not be present (the latest stable version is used then).
# Python3 + PyTest
import pytest
from appium import webdriver
# Options are available in Python client since v2.6.0
from appium.options.windows import WindowsOptions
def generate_options():
uwp_options = WindowsOptions()
# How to get the app ID for Universal Windows Apps (UWP):
# https://www.securitylearningacademy.com/mod/book/view.php?id=13829&chapterid=678
uwp_options.app = 'Microsoft.WindowsCalculator_8wekyb3d8bbwe!App'
classic_options = WindowsOptions()
classic_options.app = 'C:\\Windows\\System32\\notepad.exe'
# Make sure arguments are quoted/escaped properly if necessary:
# https://ss64.com/nt/syntax-esc.html
classic_options.app_arguments = 'D:\\log.txt'
classic_options.app_working_dir = 'D:\\'
use_existing_app_options = WindowsOptions()
# Active window handles could be retrieved from any compatible UI inspector app:
# https://docs.microsoft.com/en-us/windows/win32/winauto/inspect-objects
# or https://accessibilityinsights.io/.
# Also, it is possible to use the corresponding WinApi calls for this purpose:
# https://referencesource.microsoft.com/#System/services/monitoring/system/diagnosticts/ProcessManager.cs,db7ac68b7cb40db1
#
# This capability could be used to create a workaround for UWP apps startup:
# https://github.com/microsoft/WinAppDriver/blob/master/Samples/C%23/StickyNotesTest/StickyNotesSession.cs
use_existing_app_options.app_top_level_window = hex(12345)
return [uwp_options, classic_options, use_existing_app_options]
@pytest.fixture(params=generate_options())
def driver(request):
# The default URL is http://127.0.0.1:4723/wd/hub in Appium 1
drv = webdriver.Remote('http://127.0.0.1:4723', options=request.param)
yield drv
drv.quit()
def test_app_source_could_be_retrieved(driver):
assert len(driver.page_source) > 0
You could find more examples for different programming languages at https://github.com/microsoft/WinAppDriver/tree/master/Samples
Since version 1.15.0 of the driver there is a possibility to run custom Power Shell scripts
from your client code. This feature is potentially insecure and thus needs to be
explicitly enabled when executing the server by providing power_shell
key to the list
of enabled insecure features. Refer to Appium Security document for more details.
It is possible to ether execute a single Power Shell command (use the command
argument)
or a whole script (use the script
argument) and get its
stdout in response. If the script execution returns non-zero exit code then an exception
is going to be thrown. The exception message will contain the actual stderr.
Here's an example code of how to control the Notepad process:
// java
String psScript =
"$sig = '[DllImport(\"user32.dll\")] public static extern bool ShowWindowAsync(IntPtr hWnd, int nCmdShow);'\n" +
"Add-Type -MemberDefinition $sig -name NativeMethods -namespace Win32\n" +
"Start-Process Notepad\n" +
"$hwnd = @(Get-Process Notepad)[0].MainWindowHandle\n" +
"[Win32.NativeMethods]::ShowWindowAsync($hwnd, 2)\n" +
"[Win32.NativeMethods]::ShowWindowAsync($hwnd, 4)\n" +
"Stop-Process -Name Notepad";
driver.executeScript("powerShell", ImmutableMap.of("script", psScript));
Another example, which demonstrates how to use the command output:
# python
cmd = 'Get-Process outlook -ErrorAction SilentlyContinue'
proc_info = driver.execute_script('powerShell', {'command': cmd})
if proc_info:
print('Outlook is running')
else:
print('Outlook is not running')
Appium Windows Driver supports the same location strategies the WinAppDriver supports:
Name | Description | Speed Ranking | Example |
---|---|---|---|
accessibility id | This strategy is AutomationId attribute in inspect.exe | ⭐⭐⭐⭐⭐ |
AppNameTitle |
class name | This strategy is ClassName attribute in inspect.exe | ⭐⭐⭐⭐⭐ |
TextBlock |
id | This strategy is RuntimeId (decimal) attribute in inspect.exe | ⭐⭐⭐⭐⭐ |
42.333896.3.1 |
name | This strategy is Name attribute in inspect.exe | ⭐⭐⭐⭐⭐ |
Calculator |
tag name | This strategy is LocalizedControlType (upper camel case) attribute in inspect.exe since Appium Windows Driver 2.1.1 | ⭐⭐⭐⭐⭐ |
Text |
xpath | This strategy allows to create custom XPath queries on any attribute exposed by inspect.exe. Only XPath 1.0 is supported | ⭐⭐⭐ |
(//Button)[2] |
Beside of standard W3C APIs the driver provides the below custom command extensions to execute platform specific scenarios. Use the following source code examples in order to invoke them from your client code:
// Java 11+
var result = driver.executeScript("windows: <methodName>", Map.of(
"arg1", "value1",
"arg2", "value2"
// you may add more pairs if needed or skip providing the map completely
// if all arguments are defined as optional
));
// WebdriverIO
const result = await driver.executeScript('windows: <methodName>', [{
arg1: "value1",
arg2: "value2",
}]);
# Python
result = driver.execute_script('windows: <methodName>', {
'arg1': 'value1',
'arg2': 'value2',
})
# Ruby
result = @driver.execute_script 'windows: <methodName>', {
arg1: 'value1',
arg2: 'value2',
}
// Dotnet
object result = driver.ExecuteScript("windows: <methodName>", new Dictionary<string, object>() {
{"arg1", "value1"},
{"arg2", "value2"}
});
Record the display in background while the automated test is running. This method requires FFMPEG to be installed and present in PATH. The resulting video uses H264 codec and is ready to be played by media players built-in into web browsers.
Name | Type | Required | Description | Example |
---|---|---|---|---|
videoFilter | string | no | The video filter spec to apply for ffmpeg. See https://trac.ffmpeg.org/wiki/FilteringGuide for more details on the possible values. | Set it to scale=ifnot(gte(iw\,1024)\,iw\,1024):-2 in order to limit the video width to 1024px. The height will be adjusted automatically to match the actual ratio. |
fps | number | no | The count of frames per second in the resulting video. The greater fps it has the bigger file size is. The default vale is 15 |
10 |
preset | string | no | One of the supported encoding presets. Possible values are: ultrafast , superfast , veryfast (the default value), faster , fast , medium , slow , slower , veryslow . A preset is a collection of options that will provide a certain encoding speed to compression ratio. A slower preset will provide better compression (compression is quality per filesize). This means that, for example, if you target a certain file size or constant bit rate, you will achieve better quality with a slower preset. Read https://trac.ffmpeg.org/wiki/Encode/H.264 for more details. |
fast |
captureCursor | boolean | no | Whether to capture the mouse cursor while recording the screen. false by default |
true |
captureClicks | boolean | no | Whether to capture mouse clicks while recording the screen. false by default |
true |
timeLimit | number | no | The maximum recording time, in seconds. The default value is 600 seconds (10 minutes) | 300 |
forceRestart | boolean | no | Whether to ignore the call if a screen recording is currently running (false ) or to start a new recording immediately and terminate the existing one if running (true , the default value). |
true |
Stop recording the screen. If no screen recording has been started before then the method returns an empty string.
Name | Type | Required | Description | Example |
---|---|---|---|---|
remotePath | string | no | The path to the remote location, where the resulting video should be uploaded. The following protocols are supported: http/https, ftp. Null or empty string value (the default setting) means the content of resulting file should be encoded as Base64 and passed as the endpoint response value. An exception will be thrown if the generated media file is too big to fit into the available process memory. | https://myserver.com/upload/video.mp4 |
user | string | no | The name of the user for the remote authentication. | myname |
pass | string | no | The password for the remote authentication. | mypassword |
method | string | no | The http multipart upload method name. The 'PUT' one is used by default. | POST |
headers | map | no | Additional headers mapping for multipart http(s) uploads | {"header": "value"} |
fileFieldName | string | no | The name of the form field, where the file content BLOB should be stored for http(s) uploads. file by default |
payload |
formFields | Map or Array<Pair> |
no | Additional form fields for multipart http(s) uploads | {"field1": "value1", "field2": "value2"} or [["field1", "value1"], ["field2", "value2"]] |
Base64-encoded content of the recorded media file if remotePath
parameter is falsy or an empty string.
Remove the file from the file system. This feature is potentially insecure and thus needs to be
explicitly enabled when executing the server by providing modify_fs
key to the list
of enabled insecure features. Refer to Appium Security document for more details.
Name | Type | Required | Description | Example |
---|---|---|---|---|
remotePath | string | yes | The path to a file. The path may contain environment variables that could be expanded on the server side. Due to security reasons only variables listed below would be expanded: APPDATA , LOCALAPPDATA , PROGRAMFILES , PROGRAMFILES(X86) , PROGRAMDATA , ALLUSERSPROFILE , TEMP , TMP , HOMEPATH , USERPROFILE , PUBLIC |
%HOMEPATH%\\SomeFile.txt or C:\\Users\\user\\SomeFile.txt |
Remove the folder from the file system. This feature is potentially insecure and thus needs to
be explicitly enabled when executing the server by providing modify_fs
key to the list
of enabled insecure features. Refer to Appium Security document for more details.
Name | Type | Required | Description | Example |
---|---|---|---|---|
remotePath | string | yes | The path to a folder. The path may contain environment variables that could be expanded on the server side. Due to security reasons only variables listed below would be expanded: APPDATA , LOCALAPPDATA , PROGRAMFILES , PROGRAMFILES(X86) , PROGRAMDATA , ALLUSERSPROFILE , TEMP , TMP , HOMEPATH , USERPROFILE , PUBLIC |
%HOMEPATH%\\SomeFolder\\ or C:\\Users\\user\\SomeFolder\\ |
(Re)launch app under test in the same session using the same capabilities configuration given on the session startup. Generally this API would create a new app window and point the current active session to it, but the actual result may vary depending on how the actual application under test handles multiple instances creation. Check AppiumAppLaunch.cs for more examples. It is possible to switch between app windows using WebDriver Windows API
Close the active window of the app under test. Check AppiumAppClose.cs for more examples.
It is possible to switch between opened app windows using WebDriver Windows API.
After the current app window is closed it is required to use the above API to switch to another active window if there is any. windows: closeApp
call does not perform the switch automatically.
An error is thrown if the app under test is not running.
This is a shortcut for a single mouse click gesture.
Name | Type | Required | Description | Example |
---|---|---|---|---|
elementId | string | no | Hexadecimal identifier of the element to click on. If this parameter is missing then given coordinates will be parsed as absolute ones. Otherwise they are parsed as relative to the top left corner of this element. | 123e4567-e89b-12d3-a456-426614174000 |
x | number | no | Integer horizontal coordinate of the click point. Both x and y coordinates must be provided or none of them if elementId is present. In such case the gesture will be performed at the center point of the given element. The screen scale (if customized) is not taken into consideration while calculating the coordinate. The coordinate is always calculated for the virtual screen. | 100 |
y | number | no | Integer vertical coordinate of the click point. Both x and y coordinates must be provided or none of them if elementId is present. In such case the gesture will be performed at the center point of the given element. The screen scale (if customized) is not taken into consideration while calculating the coordinate. The coordinate is always calculated for the virtual screen. | 100 |
button | string | no | Name of the mouse button to be clicked. An exception is thrown if an unknown button name is provided. Supported button names are: left, middle, right, back, forward. The default value is left |
right |
modifierKeys | string[] or string | no | List of possible keys or a single key name to depress while the click is being performed. Supported key names are: Shift, Ctrl, Alt, Win. For example, in order to keep Ctrl+Alt depressed while clicking, provide the value of ['ctrl', 'alt'] | win |
durationMs | number | no | The number of milliseconds to wait between pressing and releasing the mouse button. By default no delay is applied, which simulates a regular click. | 500 |
times | number | no | How many times the click must be performed. One by default. | 2 |
interClickDelayMs | number | no | Duration of the pause between each click gesture. Only makes sense if times is greater than one. 100ms by default. |
10 |
This is a shortcut for a mouse wheel scroll gesture. The API is a thin wrapper over the SendInput WinApi call. It emulates the mouse cursor movement and/or horizontal/vertical rotation of the mouse wheel. Thus make sure the target control is ready to receive mouse wheel events (e.g. is focused) before invoking it.
Name | Type | Required | Description | Example |
---|---|---|---|---|
elementId | string | no | Same as in windows: click | 123e4567-e89b-12d3-a456-426614174000 |
x | number | no | Same as in windows: click | 100 |
y | number | no | Same as in windows: click | 100 |
deltaX | number | no | The amount of horizontal wheel movement measured in wheel clicks. A positive value indicates that the wheel was rotated to the right; a negative value indicates that the wheel was rotated to the left. Either this value or deltaY must be provided, but not both. | -5 |
deltaY | number | no | The amount of vertical wheel movement measured in wheel clicks. A positive value indicates that the wheel was rotated forward, away from the user; a negative value indicates that the wheel was rotated backward, toward the user. Either this value or deltaX must be provided, but not both. | 5 |
modifierKeys | string[] or string | no | Same as in windows: click | win |
This is a shortcut for a drag and drop gesture.
Name | Type | Required | Description | Example |
---|---|---|---|---|
startElementId | string | no | Same as in windows: click | 123e4567-e89b-12d3-a456-426614174000 |
startX | number | no | Same as in windows: click | 100 |
startY | number | no | Same as in windows: click | 100 |
endElementId | string | no | Same as in windows: click | 123e4567-e89b-12d3-a456-426614174000 |
endX | number | no | Same as in windows: click | 100 |
endY | number | no | Same as in windows: click | 100 |
modifierKeys | string[] or string | no | Same as in windows: click | win |
durationMs | number | no | The number of milliseconds to wait between pressing the left mouse button and moving the cursor to the ending drag point. 5000ms by default. | 7000 |
This is a shortcut for a hover gesture.
Name | Type | Required | Description | Example |
---|---|---|---|---|
startElementId | string | no | Same as in windows: click | 123e4567-e89b-12d3-a456-426614174000 |
startX | number | no | Same as in windows: click | 100 |
startY | number | no | Same as in windows: click | 100 |
endElementId | string | no | Same as in windows: click | 123e4567-e89b-12d3-a456-426614174000 |
endX | number | no | Same as in windows: click | 100 |
endY | number | no | Same as in windows: click | 100 |
modifierKeys | string[] or string | no | Same as in windows: click | win |
durationMs | number | no | The number of milliseconds between moving the cursor from the starting to the ending hover point. 500ms by default. | 700 |
This is a shortcut for a customized keyboard input.
Name | Type | Required | Description | Example |
---|---|---|---|---|
actions | KeyAction[] or KeyAction | yes | One or more KeyAction dictionaries | json [{"virtualKeyCode": 0x10, "down": true}, {'text': "appium likes you"}, {"virtualKeyCode": 0x10, "down": false}] |
Name | Type | Required | Description | Example |
---|---|---|---|---|
pause | number | no | Allows to set a delay in milliseconds between key input series. Either this property or text or virtualKeyCode must be provided. |
100 |
text | string | no | Non-empty string of Unicode text to type (surrogate characters like smileys are not supported). Either this property or pause or virtualKeyCode must be provided. |
Привіт Світ! |
virtualKeyCode | number | no | Valid virtual key code. The list of supported key codes is available at Virtual-Key Codes page. Either this property or pause or text must be provided. |
0x10 |
down | boolean | no | This property only makes sense in combination with virtualKeyCode . If set to true then the corresponding key will be depressed, false - released. By default the key is just pressed once. ! Do not forget to release depressed keys in your automated tests. |
true |
Sets Windows clipboard content to the given text or a PNG image.
Name | Type | Required | Description | Example |
---|---|---|---|---|
b64Content | string | yes | Base64-encoded content of the clipboard to be set | QXBwaXVt |
contentType | 'plaintext' or 'image' | no | Set to 'plaintext' in order to set the given text to the clipboard (the default value). Set to 'image' if b64Content contains a base64-encoded payload of a PNG image. |
image |
Retrieves Windows clipboard content.
Name | Type | Required | Description | Example |
---|---|---|---|---|
contentType | 'plaintext' or 'image' | no | Set to 'plaintext' in order to set the given text to the clipboard (the default value). Set to 'image' to retrieve a base64-encoded payload of a PNG image. | image |
Base-64 encoded content of the Windows clipboard.
Appium Windows Driver supports the following environment variables:
Variable Name | Description |
---|---|
APPIUM_WAD_PATH | A full path to WinAppDriver.exe . If you need to provide a custom path to WinAppDriver executable then set the corresponding env variable value via CMD or PowerShell: setx APPIUM_WAD_PATH "D:\New Folder\Windows Application Driver\WinAppDriver.exe" . The default location of the executable is assumed to be %PROGRAMFILES%\Windows Application Driver\WinAppDriver.exe . |
Unfortunately we cannot do much about it from the Appium Windows Drive side. It is just a thin wrapper over Microsoft's WinAppDriver closed-source REST server binary, which solely performs all the heavy lifting. This driver does handle some API calls on its own, for example PowerShell scripts execution, but the overall amount of such calls is quite limited.
Eventually your best bet would be to report the issue to WAD issues tracker and hope there is some workaround for it as Microsoft has not been regularly maintaining the driver. If it turns out the issue is just a driver regression, and it was working properly in another WAD version, then you could always replace WAD binary supplied with Appium Windows Driver by default with a custom one. Simply fetch it from Releases page and install it locally. Ideally it should transparently replace the previously installed WAD and no further actions are expected. If that did not happen though, consider providing APPIUM_WAD_PATH
environment variable pointing to the recently installed WAD binary path as described in Environment Variables section.
# Checkout the current repository and run
npm install
You can run unit and e2e tests locally:
# unit tests
npm run unit-test
# e2e tests
npm run e2e-test