Summary of an article published on The Web Scraping Club on 2026-02-01 (free to read). Read the full article

Most of us have written dozens of scripts with Selenium, Playwright or Puppeteer without looking at what sits between the library and the browser. That layer is a browser automation protocol, and it decides what your scraper can see and control. It also decides how much driver management you inherit and whether your code works outside Chromium.

I go through the three protocols that matter: WebDriver, Chrome DevTools Protocol (CDP) and WebDriver BiDi. For each one I explain what it is, which components are involved, and what happens at a low level when a script issues a command like a click or a screenshot. I also list the libraries built on each protocol and cover some history, including Selenium 3’s JSON Wire Protocol and Firefox’s remote debugging. A comparison with pros and cons closes the technical part. Then I look at where Selenium, Puppeteer, Cypress and Playwright are heading, and whether BiDi replaces CDP.

What’s inside

  • Why WebDriver needs a browser-specific driver server, and what changed with driver management in Selenium 4.10
  • How CDP organizes its commands and events into domains, and why it runs over WebSocket
  • Why Firefox and Safari have their own debug protocols, and what happened to Firefox’s limited CDP support
  • Why Puppeteer still uses CDP for Chrome even though BiDi is enabled by default for Firefox

In the full article: The step-by-step request flow for each protocol, the library tables for WebDriver and CDP, the side-by-side comparison with pros and cons, and my take on whether BiDi replaces CDP.

Read the full article on The Web Scraping Club