puppeteer/docs/webdriver-bidi.md
2024-05-29 09:14:12 +00:00

179 lines
6.2 KiB
Markdown

# Experimental WebDriver BiDi support
[WebDriver BiDi](https://w3c.github.io/webdriver-bidi/) is a new cross-browser
automation protocol currently under development, aiming to combine the best of both WebDriver “Classic” and CDP. WebDriver BiDi promises bi-directional communication, making it fast by default, and it comes packed with low-level control.
See also:
- [WebDriver BiDi - The future of cross-browser automation](https://developer.chrome.com/articles/webdriver-bidi/)
- [WebDriver BiDi: 2023 status update](https://developer.chrome.com/blog/webdriver-bidi-2023/)
## Automate with Chrome and Firefox
Puppeteer supports WebDriver BiDi automation with Chrome and Firefox.
Firefox integration is nearing feature parity with its previous CDP-based approach. Learn more in the [dedicated announcement](https://hacks.mozilla.org/2023/12/puppeteer-webdriver-bidi/).
## Measuring progress
To gauge the capabilities of WebDriver BiDi, we utilized the comprehensive [Puppeteer test suite](https://puppeteer.github.io/ispuppeteerwebdriverbidiready/)
- For Firefox, there are currently under [30](https://puppeteer.github.io/ispuppeteerwebdriverbidiready/firefox-delta.json) failing tests compared to the CDP implementation, while over [140](https://puppeteer.github.io/ispuppeteerwebdriverbidiready/firefox-delta.json) new tests successfully utilize WebDriver BiDi, demonstrating its growing potential.
- For Chrome, around 85% of tests pass with WebDriver BiDi, indicating room for improvement compared to the CDP-based approach.
## Get started
Below is an example of launching Firefox or Chrome with WebDriver BiDi:
```ts
import puppeteer from 'puppeteer';
const browser = await puppeteer.launch({
product: 'firefox', // or 'chrome'
protocol: 'webDriverBiDi',
});
const page = await browser.newPage();
...
await browser.close();
```
This is an exciting step towards a more unified and efficient cross-browser automation experience. We encourage you to explore WebDriver BiDi with Puppeteer and join us in shaping the future of browser automation.
## Puppeteer features supported over WebDriver BiDi
- Browser automation
- Puppeteer.launch
- Browser.close
- Browser.userAgent()
- Page automation
- Page.bringToFront
- Page.goBack()
- Page.goForward()
- Page.goto (except `referer` and `referrerPolicy`)
- Frame.goto() (except `referer` and `referrerPolicy`)
- Page.reload (except for `ignoreCache` parameter)
- Page.setViewport (`width`, `height`, `deviceScaleFactor` only)
- Page.cookies()
- Page.setCookie()
- Page.deleteCookie()
- Page.workers()
- PageEvent.WorkerCreated
- PageEvent.WorkerDestroyed
- Page.setExtraHTTPHeaders()
- [Script evaluation](https://pptr.dev/guides/evaluate-javascript):
- JSHandle.evaluate
- JSHandle.evaluateHandle
- Page.evaluate
- Page.exposeFunction
- [Selectors](https://pptr.dev/guides/query-selectors) and [locators](https://pptr.dev/guides/locators) except for ARIA:
- Page.$ (ARIA selectors supported in Chrome)
- Page.$$ (ARIA selectors supported in Chrome)
- Page.$$eval (ARIA selectors supported in Chrome)
- Page.$eval (ARIA selectors supported in Chrome)
- Page.waitForSelector (ARIA selectors supported in Chrome)
- Input
- ElementHandle.uploadFile
- ElementHandle.click
- Keyboard.down
- Keyboard.press
- Keyboard.sendCharacter
- Keyboard.type
- Keyboard.up
- Mouse events (except for dedicated drag'n'drop API methods)
- Page.tap
- TouchScreen.\*
- JavaScript dialog interception
- page.on('dialog')
- Dialog.\*
- Screenshots (not all parameters are supported)
- Page.screenshot (supported parameters `clip`, `encoding`, `fullPage`)
- PDF generation (not all parameters are supported)
- Page.pdf (only `format`, `height`, `landscape`, `margin`, `pageRanges`, `printBackground`, `scale`, `width` are supported)
- Page.createPDFStream (only `format`, `height`, `landscape`, `margin`, `pageRanges`, `printBackground`, `scale`, `width` are supported)
- Permissions (Supported in Chrome only)
- BrowserContext.clearPermissionOverrides()
- BrowserContext.overridePermissions()
- [Request interception](https://pptr.dev/guides/request-interception)
- HTTPRequest.abort() (no custom error support)
- HTTPRequest.abortErrorReason()
- HTTPRequest.continue()
- HTTPRequest.continueRequestOverrides()
- HTTPRequest.failure()
- HTTPRequest.finalizeInterceptions()
- HTTPRequest.interceptResolutionState()
- HTTPRequest.isInterceptResolutionHandled()
- HTTPRequest.respond()
- HTTPRequest.responseForRequest()
- Page.authenticate()
- Page.setRequestInterception()
- Page.setUserAgent()
## Puppeteer features not yet supported over WebDriver BiDi
- Various emulations (most are supported with Chrome)
- Page.emulate() (supported only in Chrome)
- Page.emulateCPUThrottling() (supported only in Chrome)
- Page.emulateIdleState() (supported only in Chrome)
- Page.emulateMediaFeatures() (supported only in Chrome)
- Page.emulateMediaType() (supported only in Chrome)
- Page.emulateTimezone() (supported only in Chrome)
- Page.emulateVisionDeficiency() (supported only in Chrome)
- Page.setBypassCSP() (supported only in Chrome)
- Page.setCacheEnabled() (supported only in Chrome)
- Page.setGeolocation() (supported only in Chrome)
- Page.setJavaScriptEnabled() (supported only in Chrome)
- CDP-specific features
- Page.createCDPSession() (supported only in Chrome)
- HTTPRequest.client() (supported only in Chrome)
- Tracing (supported only in Chrome)
- Coverage (supported only in Chrome)
- Accessibility (supported only in Chrome)
- Other methods:
- Frame.isOOPFrame()
- Frame.waitForDevicePrompt()
- HTTPResponse.buffer()
- HTTPResponse.fromServiceWorker()
- HTTPResponse.securityDetails()
- Input.drag()
- Input.dragAndDrop()
- Input.dragOver()
- Input.drop()
- Page.emulateNetworkConditions()
- Page.isDragInterceptionEnabled()
- Page.isJavaScriptEnabled() (supported only in Chrome)
- Page.isServiceWorkerBypassed()
- Page.metrics()
- Page.queryObjects() (supported only in Chrome)
- Page.screencast() (supported only in Chrome)
- Page.setBypassServiceWorker()
- Page.setDragInterception()
- Page.setOfflineMode()
- Page.waitForDevicePrompt()
- Page.waitForFileChooser()
- PageEvent.popup
- Target.opener()