In the root level of your testing folder, install the @axe-core/watcher package and all of its dependencies using npm or your preferred package manager (for example, Yarn or pnpm).

npm install --save-dev @axe-core/watcher

In your test file or files, import the puppeteerConfig function, the wrapPuppeteer() function, and the PuppeteerController class from @axe-core/watcher:

import {
  puppeteerConfig,
  wrapPuppeteerPage,
  PuppeteerController
} from '@axe-core/watcher'

Update your test setup code (typically in a before or beforeAll block) by wrapping any existing code for creating a browser instance with a call to puppeteerConfig(), while obtaining your API key from the environment (as API_KEY):

// Original code:

browser = await puppeteer.launch({
  // Your existing launch options...
})

// Becomes:

const API_KEY = process.env.API_KEY

const browser = await puppeteer.launch(
  puppeteerConfig({
    axe: {
      apiKey: API_KEY
    },
    headless: false
  })
);

Create an instance of PuppeteerController and wrap your Puppeteer browserContext:

// Create a page instance, using your browser instance.
let page = await browser.newPage()

// Initialize the PuppeteerController by passing in the Puppeteer page.
const controller = new PuppeteerController(page)

// Use the new wrapped Puppeteer page instance.
page = wrapPuppeteerPage(page, controller)

Finally, ensure all test results from axe Watcher are flushed out. To do this, add the following block of code to your test teardown code (typically in an afterEach block):

afterEach(async () => {
  await controller.flush()
});