Skip to main content


A capability is a definition for a remote interface. It helps WebdriverIO to understand in which browser or mobile environment you like to run your tests on. Capabilities are less crucial when developing tests locally as you run it on one remote interface most of the time but becomes more important when running a large set of integration tests in CI/CD.


The format of a capability object is well defined by the WebDriver specification. The WebdriverIO testrunner will fail early if user defined capabilities do not adhere to that specification.

Custom Capabilities​

While the amount of fixed defined capabilities is very low, everyone can provide and accept custom capabilities that are specific to the automation driver or remote interface:

Browser Specific Capability Extensions​

  • goog:chromeOptions: Chromedriver extensions, only applicable for testing in Chrome
  • moz:firefoxOptions: Geckodriver extensions, only applicable for testing in Firefox
  • ms:edgeOptions: EdgeOptions for specifying the environment when using EdgeDriver for testing Chromium Edge

Cloud Vendor Capability Extensions​

Automation Engine Capability Extensions​

WebdriverIO Capabilities to manage browser driver options​

WebdriverIO manages installing and running browser driver for you. WebdriverIO uses a custom capability that allows you to pass in parameters to the driver.


Specific options passed into Chromedriver when starting it.


Specific options passed into Geckodriver when starting it.


Specific options passed into Edgedriver when starting it.


Specific options passed into Safari when starting it.


Maximum number of total parallel running workers for the specific browser/capability. Takes precedence over maxInstances and maxInstancesPerCapability.

Type: number


Define specs for test execution for that browser/capability. Same as the regular specs configuration option, but specific to the browser/capability. Takes precedence over specs.

Type: (String | String[])[]


Exclude specs from test execution for that browser/capability. Same as the regular exclude configuration option, but specific to the browser/capability. Takes precedence over exclude.

Type: String[]

Common Driver Options​

While all driver offer different parameters for configuration, there are some common ones that WebdriverIO understand and uses for setting up your driver or browser:


The path to the root of the cache directory. This directory is used to store all drivers that are downloaded when attempting to start a session.

Type: string
Default: process.env.WEBDRIVER_CACHE_DIR || os.tmpdir()


Path to a custom driver binary. If set WebdriverIO won't attempt to download a driver but will use the one provided by this path. Make sure the driver is compatible with the browser you are using.

You can provide this path via CHROMEDRIVER_PATH, GECKODRIVER_PATH or EDGEDRIVER_PATH environment variables.

Type: string


If the driver binary is set, WebdriverIO won't attempt to download a driver but will use the one provided by this path. Make sure the driver is compatible with the browser you are using.

Browser Specific Driver Options​

In order to propagate options to the driver you can use the following custom capabilities:

  • Chrome or Chromium: wdio:chromedriverOptions
  • Firefox: wdio:geckodriverOptions
  • Microsoft Egde: wdio:edgedriverOptions
  • Safari: wdio:safaridriverOptions

The port on which the ADB driver should run.

Example: 9515

Type: number


Base URL path prefix for commands, e.g. wd/url.

Example: /

Type: string


Write server log to file instead of stderr, increases log level to INFO

Type: string


Set log level. Possible options ALL, DEBUG, INFO, WARNING, SEVERE, OFF.

Type: string


Log verbosely (equivalent to --log-level=ALL)

Type: boolean


Log nothing (equivalent to --log-level=OFF)

Type: boolean


Append log file instead of rewriting.

Type: boolean


Log verbosely and don't truncate long strings so that the log can be replayed (experimental).

Type: boolean


Add readable timestamps to log.

Type: boolean


Show logs from the browser (overrides other logging options).

Type: boolean


Custom bidi mapper path.

Type: string


Comma-separated allowlist of remote IP addresses which are allowed to connect to EdgeDriver.

Type: string[]
Default: ['']


Comma-separated allowlist of request origins which are allowed to connect to EdgeDriver. Using * to allow any host origin is dangerous!

Type: string[]
Default: ['*']

Special Capabilities for Specific Use Cases​

This is a list of examples showing which capabilities need to be applied to achieve a certain use case.

Run Browser Headless​

Running a headless browser means to run a browser instance without window or UI. This is mostly used within CI/CD environments where no display is used. To run a browser in headless mode, apply the following capabilities:

browserName: 'chrome', // or 'chromium'
'goog:chromeOptions': {
args: ['headless', 'disable-gpu']

Automate Different Browser Channels​

If you like to test a browser version that is not yet released as stable, e.g. Chrome Canary, you can do so by setting capabilities and pointing to the browser you like to start, e.g.:

When testing on Chrome, WebdriverIO will automatically download the desired browser version and driver for you based on the defined browserVersion, e.g.:

browserName: 'chrome', // or 'chromium'
browserVersion: '116' // or '116.0.5845.96', 'stable', 'latest', 'dev', 'canary', 'beta'

If you like to test a manually downloaded browser, you can provide a binary path to the browser via:

browserName: 'chrome', // or 'chromium'
'goog:chromeOptions': {
binary: '/Applications/Google\ Chrome\\ Chrome\ Canary'

Additionally, if you like to use a manually downloaded driver, you can provide a binary path to the driver via:

browserName: 'chrome', // or 'chromium'
'wdio:chromedriverOptions': {
binary: '/path/to/chromdriver'

Extend Custom Capabilities​

If you like to define your own set of capabilities in order to e.g. store arbitrary data to be used within the tests for that specific capability, you can do so by e.g. setting:

export const config = {
// ...
capabilities: [{
browserName: 'chrome',
'custom:caps': {
// custom configurations

It is advised to follow the W3C protocol when it comes to capability naming which requires a : (colon) character, denoting an implementation specific namespace. Within your tests you can access your custom capability through, e.g.:


In order to ensure type safety you can extend WebdriverIOs capability interface via:

declare global {
namespace WebdriverIO {
interface Capabilities {
'custom:caps': {
// ...

Welcome! How can I help?

WebdriverIO AI Copilot