Selenium - Python

         This page covers the following topics:

Installing Applitools Eyes SDK for Python

pip install eyes-selenium

Eyes Object

Eyes object is the main object in Applitools Eyes SDK.


eyes = Eyes() 


Before anything else, you should set the API key which allows you work with the Eyes server. Note that you use a static function for that, so it will be available to all of your objects. In order to obtain your API Key, login to Applitools Eyes web application and open your account details page.

eyes.api_key = "MY_API_KEY";

Replace MY_API_KEY with the API key of your account (you can find it in your signup Email).


The open method returns a custom web driver object that monitors all the driver actions. From this point, the new custom driver object should be used for all the actions in the test., app_name, test_name, viewport_size)

driver – the relevant Selenium Webdriver object.

appName – string that represents the logical name of the AUT (this name will be presented in the test result)

testName – string that represents the name of the test (this name will be presented in the test result

viewportSize – a predefined viewport size that will be used for the browser window during this test (this parameter is very important for visual tests to ensure that the test will run of a defined viewport size and the results will be consistent)

appWindow (only relevant for CodedUI) – the WinWindow or BrowserWindow object (when using BrowserWindow, a casting to WinWindow is required) of the application under test.

Match Timeout

matchTimeout defines the default amount of time (in millisecond) that Eyes will wait for an image to stabilize to a point that it is similar to the baseline image (the default value is 2 seconds, which are 2000 milliseconds).

eyes.match_timeout = TimeOut

Visual validation checkpoint

The check window command will take a snapshot of the current window in the AUT and will perform smart visual comparison of that window with the baseline. In case this is a new window that does not exist in the baseline, this window will be added to the baseline.

eyes.check_window(window_name, matchTimeout)

windowName – string that represents the logical name of this window/validation point that will appear in the test result report.

matchTimeout - (optional) matchTimeout in milliseconds, given for the step to pass.

Visual validation checkpoint of a specific region

Check region method works in a similar way to check window, except for the fact that it takes a screenshot of the HTML element specified as its input (by means of a WebDriver "By" selector) instead of a screenshot of the entire web page. The check region command will take a snapshot of the specific object, regardless of where it appears in the page, and will perform smart visual comparison of that region with the baseline. In case this is a new region that does not exist in the baseline, this region will be added as a new region to the baseline.

eyes.check_region_by_selector(selector, "window_name", matchTimeout)

selector – the “By” selector (for Selenium Webdriver) that identifies the object region. 

how - The type of the selector that will be passed in ‘what’ field. Could be one of the following: 
:class, :class_name, :css, :id, :link, :link_text, :name, :partial_link_text, :tag_name, :xpath 

what - The selection string suitable to how.

windowName – string that represents the logical name of this window/validation point that will appear in the test result report.

matchTimeout - (optional) matchTimeout in milliseconds, given for the step to pass.

Visual validation checkpoint of iframes in page

The functionality is not yet available in this language and will be added in future versions of the SDK.

Ending a test

At the end of each test, make sure to call the eyes.close method to notify the service that the test completed successfully. It is recommended to also call the method eyes.abortIfNotClosed at the end of the test as part of the ‘finally’ block to distinguish between completed tests and tests which were aborted abnormally (e.g. an exception was thrown).


In the “finally” block of Selenium tests, you would usually also want to include “driver.quit()”, since the test has either finished or failed.

(Optional) Ending tests without throwing exception on failure

The default behaviour on failure is to throw exception or raise error on test ending. To override this behaviour use the following overrides which takes a boolean.

    testResults = eyes.close(False)

testResults - Test results details, contains attributes about the test and the failure, such as:

url - A string containing the URL to the report of the test in Applitools dashboard.

steps - An integer representing the total steps count.

matches - An integer representing the total matches count.

mismatches - An integer representing the total mismatches count.

missing - An integer representing How many missing steps found.

exact_matches - An integer representing the amount of matches compared in exact match level.

strict_matches - An integer representing the amount of matches compared in strict match level.

content_Matches - An integer representing the amount of matches compared in content match level.

layout_matches - An integer representing the amount of matches compared in layout match level.

none_matches - An integer representing the amount of matches compared in None match level.

is_new - A boolean value indicates whether the test classified as a new test.

is_passed - A boolean value indicates whether the test classified as a passed test.

(Optional) Creating a batch of tests

The object BatchInfo represent a batch/collection of tests that will be unified as one group in the Test Manager screen and the result of all tests in a batch will determine the batch result. In order to start a batch, you should create a BatchInfo object, and associate it to the Eyes object (before calling the “”). In order to add additional tests to the same batch, make sure to call eyes.setBatch for each of the tests/Eyes objects (so every test in the batch should start with creating an Eyes object, calling eyes.setBatch, and ending the test with eyes.close).

batch = BatchInfo(batch_name)
eyes.batch = batch

batchName – string that represents the logical name that will be assigned to this batch.

Use the command eyes.setBatch to associate a test to the batch using the Eyes instance.

(Optional) Enabling Logs

The logHandler method of the Eyes object enables logging for troubleshooting purposes.

from applitools import logger
from applitools.logger import StdoutLogger

Export to file:

from applitools import logger
from applitools.logger import FileLogger



(Optional) Working with branches

Applitools Eyes include built-in support for branching, and allows creating separate baselines per branch, so whenever a developer makes changes in the app, he can run his visual tests locally without affecting other branches, and only commit the baseline changes after making sure that the tests are passing properly.

In order to create a branch, simply call the method eyes.setBranch after creating the Eyes object and before starting the test with the name of the branch as the parameter for this method. As a result, the subsequent tests will run as part of that branch and any baseline changes that you will perform on these tests from the web application will be private to this branch.

eyes.branch_name = branch_name

branchName – the name of the branch to create.

The initial baseline used in a branch is taken from its parent branch. You can specify the parent branch by calling the setParentBranchName method with the parent branch name as a parameter. If you don't specify a parent branch, the default branch will be used. Note that just like branching code; the baseline is copied from the parent branch only the first time it is accessed. Once the baseline is accessed in a branch, it is no longer affected by changes made in the parent branch.

eyes.parent_branch_name = parent_branch_name

parentBranchName – the name of the parent branch to inherit from.

(Optional) Merging Branches

Click here to learn how to automatically create a separate baseline per branch.

(Optional) Overriding test comparison level

Comparison sensitivity, aka ‘match level’ is set by default to Strict to get optimal coverage. Sometimes for specific tests it is required to change the default comparison level

eyes.match_level = MatchLevel.<LEVEL>

Level - The match level of the test

  • Exact - Pixel to pixel comparison, for demonstration purposes and debugging, will fail a test if a pixel is not in place. (not recommended)
  • Strict - strict is the default match level, it mimics human eyes so only significant visual changes will be spotted, while small changes that are not visible to a human eye will be ignored.
  • Content - ignores style and anti-aliasing differences, while spotting content changes. (the content level can be useful if your website includes different styles which are not relevant for your tests).
  • Layout - ignores content changes, and detects only layout changes. (useful for pages with dynamic content and localized pages)

Note: Overriding the match level is only effective if set before initialization.

(Optional) Automatic baseline creation (Default: True)

New tests presented to the server are automatically saved as baseline. To override this functionality so baseline creation will be applied manually by reviewing and approving the steps in applitools eyes Test Manager, set ‘SaveNewTests’ to false before test stat.

eyes.save_new_tests = True/False

(Optional) Auto-save on failure (Default: False)

For maintenance and debugging purposes it is possible to make eyes to save automatically failed results as baseline by setting ‘SaveFailedTests’ property to true before test started.

* It is highly unrecommended to set this property in production since all failures will be saved automatically without distinguishment between bugs and features.

eyes.save_failed_tests = True/False

(Optional) Force full page screenshot

By default when calling ‘CheckWindow’ only the visible area (viewport) will be taken in a screenshot (this is the default behavior of WebDrivers). It is possible to set Eyes to ‘stitch’ together all the scrollable areas into a single screenshot in order to ensure that the result will be a screenshot of the entire page. To set full page stitching, pass true to ‘ForceFullPageScreenshot’:

eyes.force_full_page_screenshot = True

(Optional) CSS stitching

An alternative way to perform full page screenshot (scrolling and stitching) to anchor floating UI elements (ie. top menus, help widgets) in place so they won't appear on every section that been taken.

from applitools.common import StitchMode

(Optional) Setting app environment attributes

Normally applitools eyes SDKs automatically identifies the environment that is used to run the tests and creates a separate baseline for each environment. In order to override app environment parameters and ‘force’ Eyes to compare results of different environments, you can explicitly call ‘setAppEnvironment’:

eyes.host_app = hostApp
eyes.host_os = hostOs

hostOs - The OS name and its version or any other os-alike identifier.

hostApp - The application name, if there is, such as browser name or any other app-alike identifier.

This functionality can be used to ‘force’ eyes to preset the hostOS and hostApplication so identical tests on different environments will be unified under the same baseline. Note that using this in such way will most likely cause the tests to fail, since in most cases there are some differences between how the AUT is presented in the different environments.

(Optional) Hide Scroll bars

This command hides the scrollbars while taking the screenshots.

eyes.hide_scrollbars = True

Configuring a proxy server

This section explains how to configure a proxy server in the Applitools Eyes SDK to address issues such as connection refused or connection timeout when using Applitools Eyes behind a proxy server. 

Please make sure to add the commands before calling new Eyes() and also import the os package :

import os
os.environ['HTTPS_PROXY'] ="http://localhost:8888"