Browser Automation with Node.js

You can leverage headless browser automations within Pipedream workflows for web scraping, generating screenshots, or programmatically interacting with websites - even those that make heavy usage of frontend Javascript.

Pipedream manages a specialized package (opens new window) that includes Puppeteer and Playwright bundled with a specialized Chromium instance that's compatible with Pipedream's Node.js Execution Environment.

All that's required is importing the @pipedream/browsers (opens new window) package into your Node.js code step and launch a browser. Pipedream will start Chromium and launch a Puppeteer or Playwright Browser instance for you.

• Usage
  • Puppeteer
  • Playwright
• Frequently Asked Questions
  • Can I use this package in sources or actions?
  • Workflow exited before step finished execution
  • Out of memory errors or slow starts
  • Which browser are these packages using?
  • How to customize puppeteer.launch()?

Usage

The @pipedream/browsers package exports two modules: puppeteer & playwright. Both modules share the same interface:

• browser(opts?) - method to instantiate a new browser (returns a browser instance)
• launch(opts?) - an alias to browser()
• newPage() - creates a new page instance and returns both the page & browser

Puppeteer

First import the puppeteer module from @pipedream/browsers and use browser() or launch() method to instantiate a browser.

Then using this browser you can open new Pages (opens new window), which have individual controls to open URLs:

 import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({steps, $}) {
    const browser = await puppeteer.browser();
    
    // Interact with the web page programmatically
    // See Puppeeter's Page documentation for available methods:
    // https://pptr.dev/api/puppeteer.page
    const page = await browser.newPage();

    await page.goto('https://pipedream.com/');
    const title = await page.title();
    const content = await page.content();

    $.export('title', title);
    $.export('content', content);

    // The browser needs to be closed, otherwise the step will hang
    await browser.close();
  },
})

Screenshot a webpage

Puppeteer can take a full screenshot of a webpage rendered with Chromium. For full options see the Puppeteer Screenshot method documentation. (opens new window)

Save a screenshot within the local /tmp directory:

import { puppeteer } from '@pipedreamhq/platform';

export default defineComponent({
  async run({ $ }) {
    const browser = await puppeteer.launch();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    await page.screenshot({ path: '/tmp/screenshot.png' });
    await browser.close();
  },
});

Save the screenshot as a base 64 encoded string:

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await puppeteer.browser();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com/');
    const screenshot = await page.screenshot({ encoding: 'base64' });
    await browser.close();
    return screenshot;
  },
});

Generate a PDF of a webpage

Puppeteer can render a PDF of a webpage. For full options see the Puppeteer Screenshot method documentation. (opens new window)

Save the PDF locally to /tmp:

import { puppeteer } from '@pipedreamhq/platform';

export default defineComponent({
  async run({ $ }) {
    const browser = await puppeteer.launch();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    await page.pdf({ path: '/tmp/screenshot.pdf' });
    await browser.close();
  },
});

Save the PDF as a base 64 encoded string:

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await puppeteer.browser();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    const pdfBuffer = await page.pdf();
    const pdfBase64 = pdfBuffer.toString('base64');
    await browser.close();
    return pdfBase64;
  },
});

Scrape content from a webpage

Puppeteer can scrape individual elements or return all content of a webpage.

Extract individual elements with a CSS selector:

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await puppeteer.browser();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    const h1Content = await page.$eval('h1', el => el.textContent);
    await browser.close();
    return h1Content;
  },
});

Extract all HTML content form a webpage:

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await puppeteer.browser();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    const content = await page.content();
    await browser.close();
    return content;
  },
});

Submit a form

Puppeteer can also programmatically click and type on a webpage.

Page.type accepts a CSS selector and a string to type into the field.

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({ steps, $ }) {
    const browser = await puppeteer.browser();
    const page = await browser.newPage();

    await page.goto('https://example.com/contact');
    await page.type('input[name=email]', 'pierce@pipedream.com');
    await page.type('input[name=name]', 'Dylan Pierce');
    await page.type('textarea[name=message]', "Hello, from a Pipedream workflow.");
    await page.click('input[type=submit]');

    await browser.close();
  },
});

Page.click will click on the element to focus on it, then Page.keyboard.type emulates keyboard keystrokes.

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({ steps, $ }) {
    const browser = await puppeteer.browser();
    const page = await browser.newPage();

    await page.goto('https://example.com/contact');
    await page.click('input[name=email]')
    await page.keyboard.type('pierce@pipedream.com');
    await page.click('input[name=name]')
    await page.keyboard.type('Dylan Pierce');
    await page.click('textarea[name=message]')
    await page.keyboard.type("Hello, from a Pipedream workflow.");
    await page.click('input[type=submit]');

    await browser.close();
  },
});

Playwright

First import the playwright module from @pipedream/browsers and use browser() or launch() method to instantiate a browser.

Then using this browser you can open new Pages (opens new window), which have individual controls to open URLs, click elements, generate screenshots and type and more:

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({steps, $}) {
    const browser = await playwright.browser();
    
    // Interact with the web page programmatically
    // See Playwright's Page documentation for available methods:
    // https://playwright.dev/docs/api/class-page
    const page = await browser.newPage();

    await page.goto('https://pipedream.com/');
    const title = await page.title();
    const content = await page.content();

    $.export('title', title);
    $.export('content', content);

    // The browser context and browser needs to be closed, otherwise the step will hang
    await page.context().clos);
    await browser.close();
  },
})

Don't forget to close the Browser Context

Playwright differs from Puppeteer slightly in that you have to close the page's BrowserContext before closing the Browser itself.

// Close the context & browser before returning results
// Otherwise the step will hang
await page.context().close();
await browser.close();

Screenshot a webpage

Playwright can take a full screenshot of a webpage rendered with Chromium. For full options see the Playwright Screenshot method documentation. (opens new window)

Save a screenshot within the local /tmp directory:

import { playwright } from '@pipedreamhq/platform';

export default defineComponent({
  async run({ $ }) {
    const browser = await playwright.launch();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    await page.screenshot({ path: '/tmp/screenshot.png' });

    await page.context().close();
    await browser.close();
  },
});

Save the screenshot as a base 64 encoded string:

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await playwright.launch();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com/');

    const screenshotBuffer = await page.screenshot();
    const screenshotBase64 = screenshotBuffer.toString('base64');

    await page.context().close();
    await browser.close();

    return screenshotBase64;
  },
});

Generate a PDF of a webpage

Playwright can render a PDF of a webpage. For full options see the Playwright Screenshot method documentation. (opens new window)

Save a PDF locally to /tmp:

import { playwright } from '@pipedreamhq/platform';

export default defineComponent({
  async run({ $ }) {
    const browser = await playwright.launch();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    await page.pdf({ path: '/tmp/screenshot.pdf' });

    await page.context().close();
    await browser.close();
  },
});

Save the screenshot as a base 64 encoded string:

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await playwright.launch();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com/');

    const screenshotBuffer = await page.pdf();
    const screenshotBase64 = screenshotBuffer.toString('base64');

    await page.context().close();
    await browser.close();

    return screenshotBase64;
  },
});

Scrape content from a webpage

Playwright can scrape individual elements or return all content of a webpage.

Extract individual HTML elements using a CSS Selector:

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await playwright.browser();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    const h1Content = await page.$eval('h1', el => el.textContent);

    await page.context().close();
    await browser.close();

    return h1Content;
  },
});

Extract all HTML content from a webpage with Page.content:

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({ $ }) {
    const browser = await playwright.browser();
    const page = await browser.newPage();
    await page.goto('https://pipedream.com');
    const content = await page.content();

    await page.context().close();
    await browser.close();

    return content;
  },
});

Submit a form

Playwright can also programmatically click and type on a webpage.

Page.type accepts a CSS selector and a string to type into the selected element.

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({ steps, $ }) {
    const browser = await playwright.browser();
    const page = await browser.newPage();

    await page.goto('https://example.com/contact');
    await page.type('input[name=email]', 'pierce@pipedream.com');
    await page.type('input[name=name]', 'Dylan Pierce');
    await page.type('textarea[name=message]', "Hello, from a Pipedream workflow.");
    await page.click('input[type=submit]');

    await page.context().close();
    await browser.close();
  },
});

Page.click will click on the element to focus on it, then Page.keyboard.type emulates keyboard keystrokes.

import { playwright } from '@pipedream/browsers';

export default defineComponent({
  async run({ steps, $ }) {
    const browser = await playwright.browser();
    const page = await browser.newPage();

    await page.goto('https://example.com/contact');
    await page.click('input[name=email]')
    await page.keyboard.type('pierce@pipedream.com');
    await page.click('input[name=name]')
    await page.keyboard.type('Dylan Pierce');
    await page.click('textarea[name=message]')
    await page.keyboard.type("Hello, from a Pipedream workflow.");
    await page.click('input[type=submit]');

    await page.context().close();
    await browser.close();
  },
});

Frequently Asked Questions

Can I use this package in sources or actions?

Yes, the same @pipedream/browsers package can be used in actions (opens new window) as well as sources (opens new window).

The steps are the same as usage in Node.js code. Open a browser, create a page, and close the browser at the end of the code step.

Memory limits

At this time it's not possible to configure the allocated memory to a Source. You may experience a higher rate of Out of Memory errors on Sources that use Puppeteer or Playwright due to the high usage of memory required by Chromium.

Workflow exited before step finished execution

Remember to close the browser instance before the step finishes. Otherwise, the browser will keep the step "open" and not transfer control to the next step.

Out of memory errors or slow starts

For best results, we recommend increasing the amount of memory available to your workflow to 2 gigabytes. You can adjust the available memory in the workflow settings (opens new window).

Which browser are these packages using?

The @pipedream/browsers package includes a specific version of Chromium that is compatible with Pipedream Node.js execution environments that run your code.

For details on the specific versions of Chromium, puppeeter and playwright bundled in this package, visit the package's README (opens new window).

How to customize puppeteer.launch()?

To pass arguments to puppeteer.launch() to customize the browser instance, you can pass them directly to puppeteer.browser().

For example, you can alter the protocolTimeout length just by passing it as an argument:

import { puppeteer } from '@pipedream/browsers';

export default defineComponent({
  async run({steps, $}) {
    // extend the maximum timeout length for puppeteer operations:
    const browser = await puppeteer.browser({ protocolTimeout: 480000 });
  },
})

See the source code here (opens new window) for the default arguments that Pipedream provides.