CodeNode.jsBrowser automation

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 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 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

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, 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.

Save a screenshot within the local /tmp directory:

import { puppeteer } from '@pipedream/browsers';
 
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();
  },
});

Generate a PDF of a webpage

Puppeteer can render a PDF of a webpage. For full options see the Puppeteer Screenshot method documentation.

Save the PDF locally to /tmp:

import { puppeteer } from '@pipedream/browsers';
 
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();
  },
});

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;
  },
});

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();
  },
});

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, 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().close();
    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 browser context 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.

Save a screenshot within the local /tmp directory:

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');
    await page.screenshot({ path: '/tmp/screenshot.png' });
 
    await page.context().close();
    await browser.close();
  },
});

Generate a PDF of a webpage

Playwright can render a PDF of a webpage. For full options see the Playwright Screenshot method documentation.

Save a PDF locally to /tmp:

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');
    await page.pdf({ path: '/tmp/screenshot.pdf' });
 
    await page.context().close();
    await browser.close();
  },
});

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;
  },
});

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();
  },
});

Frequently Asked Questions

Can I use this package in sources or actions?

Yes, the same @pipedream/browsers package can be used in actions as well as sources.

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.

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.

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, $}) {
    // passing a `protocolTimeout` argument to increase the timeout length for a puppeteer instance
    const browser = await puppeteer.browser({ protocolTimeout: 480000 });
    // rest of code
  },
})

Please see the @pipedream/browsers source code for the default arguments that Pipedream provides.

How do I use puppeteer.connect()?

To use puppeteer.connect() to connect to a remote browser instance, you can use the puppeteer-core package:

import puppeteer from "puppeteer-core";

puppeteer-core does not download Chrome when installed, which decreases the size of your deployment and can improve cold start times.

To connect to a remote browser instance using Playwright, you can use the playwright-core package, which is the no-browser Playwright package:

import playwright from "playwright-core";
Sharing code across workflowsRunning Python in workflows