Skip to main content

Setup Types

WebdriverIO can be used for various purposes. It implements the WebDriver protocol API and can run a browser in an automated way. The framework is designed to work in any arbitrary environment and for any kind of task. It is independent from any 3rd party frameworks and only requires Node.js to run.

Protocol Bindings

For basic interactions with the WebDriver and other automation protocols WebdriverIO uses its own protocol bindings based on the webdriver NPM package:


All protocol commands return the raw response from the automation driver. The package is very lightweight and there is no smart logic like auto-waits to simplify the interaction with the protocol usage.

The protocol commands applied to the instance depend on the initial session response of the driver. For example if the response indicates that a mobile session was started, the package applies all Appium and Mobile JSON Wire protocol commands to the instance prototype.

You can run the same set of commands (except mobile ones) using the Chrome DevTools protocol when importing the devtools NPM package. It has the same interface as the webdriver package but runs its automation based on Puppeteer.

For more information on these package interfaces, see Modules API.

Standalone Mode

To simplify the interaction with the WebDriver protocol the webdriverio package implements a variety of commands on top of the protocol (e.g. the dragAndDrop command) and core concepts such as smart selectors or auto-waits. The example from above can be simplified like this:


Using WebdriverIO in standalone mode still gives you access to all protocol commands but provides a super set of additional commands that provide a higher level interaction with the browser. It allows you to integrate this automation tool in your own (test) project to create a new automation library. Popular examples include Spectron or CodeceptJS. You can also write plain Node scripts to scrape the web for content (or anything else that requires a running browser).

If no specific options are set WebdriverIO will always attempt to download and setup the browser driver that matches browserName property in your capabilities. In case of Chrome and Firefox it might also install them depending on whether it can find the corresponding browser on the machine.

For more information on the webdriverio package interfaces, see Modules API.

The WDIO Testrunner

The main purpose of WebdriverIO, though, is end-to-end testing on a big scale. We therefore implemented a test runner that helps you to build a reliable test suite that is easy to read and maintain.

The test runner takes care of many problems that are common when working with plain automation libraries. For one, it organizes your test runs and splits up test specs so your tests can be executed with maximum concurrency. It also handles session management and provides lots of features to help you to debug problems and find errors in your tests.

Here is the same example from above, written as a test spec and executed by WDIO:


The test runner is an abstraction of popular test frameworks like Mocha, Jasmine, or Cucumber. To run your tests using the WDIO test runner, check out the Getting Started section for more information.

For more information on the @wdio/cli testrunner package interface, see Modules API.

Welcome! How can I help?

WebdriverIO AI Copilot