Multiple Local Testing connections

If you are using same account to test multiple applications, you can setup a named connection using localIdentifier.

1. Run the BrowserStackLocal binary with the localIdentifier option

   • OS X & Linux
   • Windows

   BrowserStackLocal.exe --key ACCESS_KEY --local-identifier Test123

   Here, we have created a named connection called Test123.

2. Once the connection is made, you need to set the browserstack.localIdentifier capability with the correct named connection

   caps['browserstack.local'] = 'true'
   caps['browserstack.localIdentifier'] = 'Test123'
   

Run tests on mobile browsers

You can now run your Selenium and Appium test scripts on real Android and iOS devices. Use the drop down below to select your device.
Look for the icon to select a real device.

caps[:browserName] = 'iPhone'
caps[:platform] = 'MAC'
caps['device'] = 'iPhone 5'

For a list of all supported devices, visit the Browsers and Platforms page.

Builds and projects

Keep track of all your automated tests using the build and project capabilities. Group your tests into builds, and builds further into projects.

caps['build'] = 'version1'
caps['project'] = 'newintropage'

Self-signed certificates

To avoid invalid certificate errors while testing on BrowserStack Automate, set the acceptSslCerts capability in your test to true.

caps['acceptSslCerts'] = 'true'

Enable and Disable Flash

To disable Flash in Chrome, create a chromeOptions capability, and pass the --disable-plugins argument to the capability.

caps = Selenium::WebDriver::Remote::Capabilities.chrome
caps["chromeOptions"] = {}
caps["chromeOptions"]["args"] = ["--disable-plugins"]

It is possible to disable Flash within Firefox, by setting the profile capability to 0. Flash is enabled by default.

profile = Selenium::WebDriver::Firefox::Profile.new
profile["plugin.state.flash"] = 0
caps = Selenium::WebDriver::Remote::Capabilities.firefox(:firefox_profile => profile)

To disable Flash in Internet explorer, pass the browserstack.ie.noFlash capability in your tests.

caps["browserstack.ie.noFlash"] = "true"

Enable and Disable Pop-ups

Popup blocker is disabled by default in Chrome >= 43. To enable the popup blocker, create a chromeOptions capability, and pass the disable-popup-blocking as excludeSwitches to the capability.

caps = Selenium::WebDriver::Remote::Capabilities.chrome
caps["chromeOptions"] = {}
caps["chromeOptions"]["excludeSwitches"] = ["disable-popup-blocking"]

To disable the popup blocker in Chrome < 43, create a chromeOptions capability, and pass the --disable-popup-blocking argument to the capability.

caps = Selenium::WebDriver::Remote::Capabilities.chrome
caps["chromeOptions"] = {}
caps["chromeOptions"]["args"] = ["--disable-popup-blocking"]

To enable the popups in IE, use the browserstack.ie.enablePopups capability.

caps['browserstack.ie.enablePopups'] = 'true'

To enable the popups in Safari, use the browserstack.safari.enablePopups capability.

caps['browserstack.safari.enablePopups'] = 'true'

Debugging

To debug failed tests, BrowserStack provides you with raw logs, which are the console logs from the browser tests; visual logs, which capture successive screenshots of the test; and text logs to display all the steps that were performed by the test.

With live screencast, view ongoing tests to debug the functionality of features that are being tested. The generated videos can be recorded and downloaded for later viewing.

Logs and the live screencast help you to compare and detect any changes that may have occurred since a similar test was last run. Debugging is set to false by default. To enable logs, set the BrowserStack custom capability browserstack.debug to true.

caps['browserstack.debug'] = 'true'

Videos are recorded for every BrowserStack Automate test. This feature helps you to verify and debug failed tests, confirm the functionality of tested features, and download videos for later viewing.

caps['browserstack.video'] = 'false'

Other capabilities

BrowserStack supports the full complement of Selenium capabilities, as well as, some custom ones.

ChromeOptions

A ChromeDriver session can further be customized and configured using ChromeOptions. BrowserStack supports the full complement of ChromeOptions. In this example, ChromeOptions is used to disable flash in Chrome.

caps = Selenium::WebDriver::Remote::Capabilities.chrome
caps["chromeOptions"] = {}
caps["chromeOptions"]["args"] = ["--disable-plugins"]

Firefox Profile

To change the preferences within Firefox, instantiate a new Firefox Profile object and update the preferences of the profile. Check out more preferences at the MozillaZine Knowledge Base.

In this example, a custom Firefox Profile is used to disable flash in Firefox.

profile = Selenium::WebDriver::Firefox::Profile.new
profile["plugin.state.flash"] = 0
caps = Selenium::WebDriver::Remote::Capabilities.firefox(:firefox_profile => profile)

Taking screenshots

To take screenshots automatically from within the tests:

driver.save_screenshot("screenshots.png")

Basic HTTP Authentication

If the website uses basic authentication, use the code snippet below as a reference to connect to the website via your Selenium tests

driver.navigate.to("http://<username>:<password>@yourdomain")

Uploads & Downloads

To upload files through tests:

driver.file_detector = lambda do |args|
  str = args.first.to_s
  str if File.exist?(str)
end
driver.navigate.to "http://www.fileconvoy.com"
driver.find_element(:id, "upfile_0").send_keys("C:\\Users\\hello\\url.txt");
driver.find_element(:id, "readTermsOfUse").click;
driver.find_element(:name, "form_upload").submit;
sleep(5)

To download files through the test files, you need to create a profile capability containing all the necessary parameters, and then associate it with the WebDriver.

profile = Selenium::WebDriver::Firefox::Profile.new
profile['browser.download.folderList'] = 0
profile['browser.download.manager.showWhenStarting'] = false
profile['browser.download.manager.focusWhenStarting'] = false
profile['browser.download.useDownloadDir'] = true
profile['browser.helperApps.alwaysAsk.force'] = false
profile['browser.download.manager.alertOnEXEOpen'] = false
profile['browser.download.manager.closeWhenDone'] = true
profile['browser.download.manager.showAlertOnComplete'] = false
profile['browser.download.manager.useWindow'] = false
# you will need to find the content-type of your app and set it here.
profile['browser.helperApps.neverAsk.saveToDisk'] = "application/octet-stream"
caps = Selenium::WebDriver::Remote::Capabilities.firefox(:firefox_profile => profile)
driver = Selenium::WebDriver.for(:remote,
                                 :url => "http://USERNAME:ACCESS_KEY@hub-cloud.browserstack.com/wd/hub",
                                 :desired_capabilities => caps)
driver.navigate.to "https://rubygems.org/gems/selenium-webdriver"
driver.find_element(:id => 'download').click
driver.quit

Session ID

To obtain the session ID of a running BrowserStack Automate session:

@driver.session_id

Running tests from behind a proxy

To run your tests from behind your local proxy:

@driver.proxy = Selenium::Proxy.new(:http => "proxy.org:8080")

REST API

It is possible to mark tests as either a pass or a fail, using the following snippet:

require 'rest-client'
RestClient.put "https://USERNAME:ACCESS_KEY@www.browserstack.com/automate/sessions/<session-id>.json", {"status"=>"passed", "reason"=>""}, {:content_type => :json}

The two potential values for status can either be passed or failed. Optionally, a reason can also be passed.

Parallel tests

Parallel testing is running multiple tests concurrently to reduce your testing time, and is a key feature of BrowserStack Automate. This can be the same test running on different browser configurations, or different tests running on the same browser configuration. With access to our extensive BrowserStack cloud, parallel testing on many platforms becomes very efficient.

Queuing

With queuing, you can launch an additional number of parallel tests with different browser configurations that will be queued in a sequence, for a seamless parallel execution. For instance, if you want to run 5 additional tests, apart from your subscribed limit of 2 parallel tests, BrowserStack will queue the additional 5 tests until one of the 2 initial tests finish, and a slot is available for execution. With queuing, you can be less bothered about managing your tests, and it can save development time.

With this feature, accounts up to 5 parallel tests can queue 5 tests. Beyond 5 parallel tests, an equivalent number of tests will be queued.

We have provided examples of parallel testing implementation using popular testing frameworks. In order to increase the number of tests, purchase more parallel tests of your Automate or Automate Pro plan to get access to more tests.

Add-on

We provide a ‘fast-selenium’ add-on to speed up your tests even more. All you have to do is download it, and include it in your code, with a require statement:

require './fast-selenium'

Test::Unit

In the following test, the username and access key are inputted via the test.

require 'rubygems'
require 'selenium-webdriver'
require 'test/unit'

class SampleTest < Test::Unit::TestCase
  def setup
    url = "http://USERNAME:ACCESS_KEY@hub-cloud.browserstack.com/wd/hub"
    @driver = Selenium::WebDriver.for(:remote, :url => url)
  end

  def test_post
    @driver.navigate.to "http://www.google.com"
    element = @driver.find_element(:name, 'q')
    element.send_keys "BrowserStack"
    element.submit
    assert_equal(@driver.title, "BrowserStack - Google Search")
  end

  def teardown
    @driver.quit
  end
end

Add the parallel testing gem to the project Gemfile, group all your tests together, and run the command. Each of the tests has to be configured to run remotely, as described above.

The example below uses the Test::Unit framework to run parallel tests, using two tests. The username and access key are set as environment variables within the code, which when passed as parameters to the program, they are then accessible throughout the testing framework.

Test 1: google-search-browserstack-chrome-mac.rb

require 'rubygems'
require 'selenium-webdriver'
require 'test/unit'
 
class SampleTest1 < Test::Unit::TestCase
  def setup
    username=ENV['BS_USERNAME']
    key=ENV['BS_AUTHKEY']
    url = "http://#{username}:#{key}@hub-cloud.browserstack.com/wd/hub"
    capabilities = Selenium::WebDriver::Remote::Capabilities.chrome
    capabilities.platform = :MAC
    @driver = Selenium::WebDriver.for(:remote, :url => url,  :desired_capabilities => capabilities)
  end
 
  def test_post
    @driver.navigate.to "http://www.google.com"
    element = @driver.find_element(:name, 'q')
    element.send_keys "BrowserStack"
    element.submit
    assert_equal(@driver.title, "BrowserStack - Google Search")
  end
 
  def teardown
    @driver.quit
  end
end

Test 2: google-search-browserstack-ie9.rb

require 'rubygems'
require 'selenium-webdriver'
require 'test/unit'
 
class SampleTest2 < Test::Unit::TestCase
  def setup
    username=ENV['BS_USERNAME']
    key=ENV['BS_AUTHKEY']
    url = "http://#{username}:#{key}@hub-cloud.browserstack.com/wd/hub"
    capabilities = Selenium::WebDriver::Remote::Capabilities.internet_explorer
    capabilities.version = "9.0"
    capabilities.platform = :WINDOWS
    @driver = Selenium::WebDriver.for(:remote, :url => url,  :desired_capabilities => capabilities)
  end
 
  def test_post
    @driver.navigate.to "http://www.google.com"
    element = @driver.find_element(:name, 'q')
    element.send_keys "BrowserStack"
    element.submit
    assert_equal(@driver.title, "BrowserStack - Google Search")
  end
 
  def teardown
    @driver.quit
  end
end

parallel_test -n 2 test/google-search-browserstack-ie9.rb test/google-search-browserstack-chrome-mac.rb

Cucumber

To start testing using Cucumber, you have to install the relevant dependencies.

  source 'http://rubygems.org'
  gem 'cucumber'
  gem 'selenium-webdriver'
  gem 'rspec'

Next, define a features and a steps file, which together constitute your test.

Feature: Google can search

Background:
  Given I am on Google

Scenario: Search for a term
  When I fill in "q" found by "name" with "BrowserStack"
  And I submit
  Then I should see title "BrowserStack - Google Search"

Given /^I am on (.+)$/ do |url|
  @browser.navigate.to "http://www.google.com"
end

When /^I fill in "([^"]*)" found by "([^"]*)" with "([^"]*)"$/ do |value, type, keys|
  @element = @browser.find_element(type, value)
  @element.send_keys keys
end

When /^I submit$/ do
  @element.submit
end

Then /^I should see title "([^"]*)"$/ do |title|
  raise "Fail" if @browser.title != title
end

To get Cucumber to use the BrowserStack infrastructure, you will need to configure its capabilities accordingly, in the environment variable file, i.e. features/support/env.rb.

require 'selenium/webdriver'

url = "http://#{ENV['BS_USERNAME']}:#{ENV['BS_AUTHKEY']}@hub-cloud.browserstack.com/wd/hub"

capabilities = Selenium::WebDriver::Remote::Capabilities.new

capabilities['project'] = ENV['BS_AUTOMATE_PROJECT'] if ENV['BS_AUTOMATE_PROJECT']
capabilities['build'] = ENV['BS_AUTOMATE_BUILD'] if ENV['BS_AUTOMATE_BUILD']

if ENV['BS_AUTOMATE_OS']
  capabilities['os'] = ENV['BS_AUTOMATE_OS']
  capabilities['os_version'] = ENV['BS_AUTOMATE_OS_VERSION']
else
  capabilities['platform'] = ENV['SELENIUM_PLATFORM'] || 'ANY'
end

capabilities['browser'] = ENV['SELENIUM_BROWSER'] || 'chrome'
capabilities['browser_version'] = ENV['SELENIUM_VERSION'] if ENV['SELENIUM_VERSION']

browser = Selenium::WebDriver.for(:remote, :url => url, :desired_capabilities => capabilities)

Before do |scenario|
  @browser = browser
end

at_exit do
  browser.quit
end

Finally, run the test from the command-line interface, passing the parameters for the test environment.

cucumber BS_USERNAME=USERNAME BS_AUTHKEY=ACCESS_KEY SELENIUM_PLATFORM=WINDOWS SELENIUM_BROWSER=chrome

To start testing using Cucumber, you have to install the relevant dependencies.

source 'http://rubygems.org'
gem 'cucumber'
gem 'selenium-webdriver'
gem 'rspec'
gem 'parallel'
      

The browsers.json file defines the combinations on which the tests will be run.

[
  {
    "browser": "firefox",
    "browser_version": "15.0",
    "device": null,
    "os": "OS X",
    "os_version": "Snow Leopard"
  },
  {
    "browser": "firefox",
    "browser_version": "16.0",
    "device": null,
    "os": "Windows",
    "os_version": "7"
  },
  {
    "browser": "firefox",
    "browser_version": "17.0",
    "device": null,
    "os": "Windows",
    "os_version": "8"
  },
  {
    "browser": "ie",
    "browser_version": "8.0",
    "device": null,
    "os": "Windows",
    "os_version": "7"
  },
  {
    "browser": "ie",
    "browser_version": "9.0",
    "device": null,
    "os": "Windows",
    "os_version": "7"
  },
  {
    "browser": "ie",
    "browser_version": "10.0",
    "device": null,
    "os": "Windows",
    "os_version": "8"
  }
]

Next, define a features and a steps file, which together constitute your tests.

Feature: Google can search

Background:
  Given I am on Google

Scenario: Search for a term
  When I fill in "q" found by "name" with "BrowserStack"
  And I submit
  Then I should see title "BrowserStack - Google Search"

Given /^I am on (.+)$/ do |url|
  @browser.navigate.to "http://www.google.com"
end

When /^I fill in "([^"]*)" found by "([^"]*)" with "([^"]*)"$/ do |value, type, keys|
  @element = @browser.find_element(type, value)
  @element.send_keys keys
end

When /^I submit$/ do
  @element.submit
end

Then /^I should see title "([^"]*)"$/ do |title|
  raise "Fail" if @browser.title != title
end

Setup the username and access key environment variables, and the capabilities, to define the parameters for all the tests. Values will be passed to this via the command-line interface.

require 'selenium/webdriver'

url = "http://#{ENV['BS_USERNAME']}:#{ENV['BS_AUTHKEY']}@hub-cloud.browserstack.com/wd/hub"

capabilities = Selenium::WebDriver::Remote::Capabilities.new
capabilities['os'] = ENV['BS_AUTOMATE_OS']
capabilities['os_version'] = ENV['BS_AUTOMATE_OS_VERSION']
capabilities['browser'] = ENV['SELENIUM_BROWSER']
capabilities['browser_version'] = ENV['SELENIUM_VERSION']
capabilities['browserstack.debug'] = "true"

capabilities['project'] = ENV['BS_AUTOMATE_PROJECT'] if ENV['BS_AUTOMATE_PROJECT']
capabilities['build'] = ENV['BS_AUTOMATE_BUILD'] if ENV['BS_AUTOMATE_BUILD']

browser = Selenium::WebDriver.for(:remote, :url => url,
                                  :desired_capabilities => capabilities)

Before do |scenario|
  @browser = browser
end

at_exit do
  browser.quit
end

Create a Rakefile, which handles test management. If there are only three nodes available, for example, and the total number of tests exceeds that, then the Rakefile will schedule the tests to operate within those constraints.

require 'rubygems'
require 'cucumber'
require 'cucumber/rake/task'
require 'parallel'
require 'json'

@browsers = JSON.load(open('browsers.json'))
@parallel_limit = ENV["nodes"] || 1
@parallel_limit = @parallel_limit.to_i

task :cucumber do
  Parallel.each(@browsers, :in_processes => @parallel_limit) do |browser|
    begin
      puts "Running with: #{browser.inspect}"
      ENV['SELENIUM_BROWSER'] = browser['browser']
      ENV['SELENIUM_VERSION'] = browser['browser_version']
      ENV['BS_AUTOMATE_OS'] = browser['os']
      ENV['BS_AUTOMATE_OS_VERSION'] = browser['os_version']

      Rake::Task[:run_features].execute()
    rescue Exception => e
      puts "Error while running task"
    end
  end
end

Cucumber::Rake::Task.new(:run_features)
task :default => [:cucumber]

Finally, run your tests by passing the username, access key, and number of parallel tests via the command-line interface.

rake BS_USERNAME=USERNAME BS_AUTHKEY=ACCESS_KEY nodes=3

Minitest

require 'rubygems'
require 'minitest/autorun'
require 'selenium-webdriver'

class GoogleTest < MiniTest::Test
  USERNAME = ''
  BROWSERSTACK_ACCESS_KEY = ''
  def setup
    if USERNAME == ''
      puts "Please add USERNAME & BROWSERSTACK_ACCESS_KEY in this file"
      exit
    end
    url = "http://#{USERNAME}:#{BROWSERSTACK_ACCESS_KEY}@hub-cloud.browserstack.com/wd/hub"
    @driver = Selenium::WebDriver.for(:remote, :url => url)
  end

  def test_post
    @driver.navigate.to "http://www.google.com"
    element = @driver.find_element(:name, 'q')
    element.send_keys "BrowserStack"
    element.submit
    assert_equal(@driver.title, "BrowserStack - Google Search")
  end

  def teardown
    @driver.quit
  end
end

The browsers.json file defines the combinations on which the tests will be run.

[
  {
    "browser": "firefox",
    "browser_version": "15.0",
    "device": null,
    "os": "OS X",
    "os_version": "Snow Leopard"
  },
  {
    "browser": "firefox",
    "browser_version": "16.0",
    "device": null,
    "os": "Windows",
    "os_version": "7"
  },
  {
    "browser": "firefox",
    "browser_version": "17.0",
    "device": null,
    "os": "Windows",
    "os_version": "8"
  },
  {
    "browser": "ie",
    "browser_version": "8.0",
    "device": null,
    "os": "Windows",
    "os_version": "7"
  },
  {
    "browser": "ie",
    "browser_version": "9.0",
    "device": null,
    "os": "Windows",
    "os_version": "7"
  },
  {
    "browser": "ie",
    "browser_version": "10.0",
    "device": null,
    "os": "Windows",
    "os_version": "8"
  }
]

Setup the username and access key environment variables, and the capabilities, to define the parameters for all the tests. Values will be passed to this via the command-line interface.

require 'rubygems'
require 'minitest/autorun'
require 'selenium-webdriver'

class GoogleTest < MiniTest::Test
  USERNAME = ENV['BS_USERNAME']
  BROWSERSTACK_ACCESS_KEY = ENV['BS_AUTHKEY']
  
  def setup
    if USERNAME == ''
      puts "Please add username & key as parameters while running rake task"
      exit
    end
    url = "http://#{USERNAME}:#{BROWSERSTACK_ACCESS_KEY}@hub-cloud.browserstack.com/wd/hub"
    capabilities = Selenium::WebDriver::Remote::Capabilities.new
    capabilities['os'] = ENV['BS_AUTOMATE_OS']
    capabilities['os_version'] = ENV['BS_AUTOMATE_OS_VERSION']
    capabilities['browser'] = ENV['SELENIUM_BROWSER']
    capabilities['browser_version'] = ENV['SELENIUM_VERSION']
    @driver = Selenium::WebDriver.for(:remote,
                                      :url => url,
                                      :desired_capabilities => capabilities)
  end
 
  def test_post
    @driver.navigate.to "http://www.google.com"
    element = @driver.find_element(:name, 'q')
    element.send_keys "BrowserStack"
    element.submit
    assert_equal(@driver.title, "BrowserStack - Google Search")
  end
 
  def teardown
    @driver.quit
  end
end

Create a Rakefile, which handles test management. If there are only three nodes available, for example, and the total number of tests exceeds that, then the Rakefile will schedule the tests to operate within those constraints.

require 'rubygems'
require 'rake/testtask'
require 'parallel'
require 'json'

@browsers = JSON.load(open('browsers.json'))
@test_folder = "test/*_test.rb"
@parallel_limit = ENV["nodes"] || 1
@parallel_limit = @parallel_limit.to_i

task :minitest do
  current_browser = ""
  begin
    Parallel.map(@browsers, :in_threads => @parallel_limit) do |browser|
      current_browser = browser
      puts "Running with: #{browser.inspect}"
      ENV['SELENIUM_BROWSER'] = browser['browser']
      ENV['SELENIUM_VERSION'] = browser['browser_version']
      ENV['BS_AUTOMATE_OS'] = browser['os']
      ENV['BS_AUTOMATE_OS_VERSION'] = browser['os_version']
      Dir.glob(@test_folder).each do |test_file|
        IO.popen("ruby #{test_file}") do |io|
          io.each do |line|
            puts line
          end
        end
      end
    end
  rescue SystemExit, Interrupt
    puts "User stopped script!"
    puts "Failed to run tests for #{current_browser.inspect}"
  end
end

task :default => [:minitest]

Finally, run your tests by passing the username, access key, and number of parallel tests via the command-line interface.

rake BS_USERNAME=USERNAME BS_AUTHKEY=ACCESS_KEY nodes=3

Watir

To use Watir on BrowserStack, the tests must be configured to run on remote virtual machines. This is easily achieved by changing the capabilities of the webdriver. For instance, the following code would direct the webdriver to automate tests for Firefox on your local machine.

browser = Watir::Browser.new(:firefox)

Therefore, for the code to run on BrowserStack, the Watir capabilities have to be changed.

browser = Watir::Browser.new(:remote,
  :url => "http://USERNAME:ACCESS_KEY@hub-cloud.browserstack.com/wd/hub",
  :desired_capabilities => caps)

The example is a simple test made in Watir to run on Automate. It opens Google's homepage, searches for ‘browserstack’, and asks for the title of the search results page.

require 'rubygems'
require 'watir-webdriver'

include Selenium

caps = WebDriver::Remote::Capabilities.new
caps[:os] = "Windows"
caps[:name] = "Watir WebDriver"
caps[:browser] = "firefox"
caps[:browser_version] = "20"
caps["browserstack.debug"] = "true"

browser = Watir::Browser.new(:remote,
  :url => "http://USERNAME:ACCESS_KEY@hub-cloud.browserstack.com/wd/hub",
  :desired_capabilities => caps)

browser.goto "http://www.google.com"
browser.text_field(:name => 'q').set 'BrowserStack'
browser.button(:name => 'btnG').click

puts browser.title
browser.quit

Capybara

To integrate Capybara with BrowserStack infrastructure to run Cucumber tests, you will need to configure capabilities in the environment variable file.

The following example configures the capabilities in features/support/env.rb:

#features/sample.feature

Feature: Google can search

Background:
  Given I am on Google

Scenario: Search for a term
  When I fill in "q" found by "name" with "BrowserStack"
  And I submit
  Then I should see title "BrowserStack - Google Search"

#features/sample_steps.rb

Given /^I am on (.+)$/ do |url|
  visit "/"
end
 
When /^I fill in "([^\"]*)" found by "([^\"]*)" with "([^\"]*)"$/ do |value, type, keys|
  fill_in(value, :with => keys)
end
 
When /^I submit$/ do
  find_field('q').native.send_key(:enter)
end
 
Then /^I should see title "([^\"]*)"$/ do |title|
  expect(page).to have_title title
end

#features/support/env.rb

require 'selenium/webdriver'
require 'capybara/cucumber'

url = "http://#{ENV['BS_USERNAME']}:#{ENV['BS_AUTHKEY']}@hub-cloud.browserstack.com/wd/hub"

Capybara.register_driver :browserstack do |app|
    capabilities = Selenium::WebDriver::Remote::Capabilities.new

    if ENV['BS_AUTOMATE_OS']
        capabilities['os'] = ENV['BS_AUTOMATE_OS']
        capabilities['os_version'] = ENV['BS_AUTOMATE_OS_VERSION']
    else
        capabilities['platform'] = ENV['SELENIUM_PLATFORM'] || 'ANY'
    end

    capabilities['browser'] = ENV['SELENIUM_BROWSER'] || 'chrome'
    capabilities['browser_version'] = ENV['SELENIUM_VERSION'] if ENV['SELENIUM_VERSION']


    capabilities['browserstack.debug'] = 'true'
    capabilities['project'] = ENV['BS_AUTOMATE_PROJECT'] if ENV['BS_AUTOMATE_PROJECT']
    capabilities['build'] = ENV['BS_AUTOMATE_BUILD'] if ENV['BS_AUTOMATE_BUILD']

    Capybara::Selenium::Driver.new(app,
                                 :browser => :remote, :url => url,
                                 :desired_capabilities => capabilities)
end


Capybara.default_driver = :browserstack
Capybara.app_host = "http://www.google.com"
Capybara.run_server = false

Finally, run the test from the command-line interface, passing the parameters for the test environment.

cucumber BS_USERNAME=USERNAME BS_AUTHKEY=ACCESS_KEY SELENIUM_PLATFORM=WINDOWS SELENIUM_BROWSER=chrome