# What is the Chrome DevTools Protocol (CDP) in web scraping?

The Chrome DevTools Protocol (CDP) is a set of APIs that exposes a running Chromium browser over a WebSocket connection, letting external tools open tabs, evaluate JavaScript, intercept network requests, and capture screenshots programmatically. Libraries like [Playwright](/content/glossary/web-scraping-apis/playwright-web-scraping/index.html) and Puppeteer use CDP under the hood to automate browsers for scraping. When you call `page.goto()` or `page.screenshot()`, those calls translate to CDP commands sent over a WebSocket to the browser process.

| Factor                     | Playwright / Puppeteer | Direct CDP               |
|---------------------------|-----------------------|--------------------------|
| Abstraction level         | High-level API        | Raw protocol commands     |
| Setup                     | Library install       | WebSocket client only    |
| Use case                  | Standard scraping automation | Custom tooling, session attachment |
| Debugging                 | Built-in error handling | Full low-level visibility |
| Learning curve            | Low                   | High                     |

Use CDP directly when you need to attach your own tooling to a browser session that is already running, intercept and modify requests at the protocol level, or connect Playwright to a remote browser managed by a third-party service. For most scraping tasks, [Playwright](/content/glossary/web-scraping-apis/playwright-web-scraping/index.html) or a scraping API is faster to work with. Direct CDP access becomes relevant for custom session takeover, low-level traffic inspection, or hybrid setups where one service manages the browser and another controls it.

[Firecrawl Browser Sandbox](https://docs.firecrawl.dev/features/browser) exposes a live WebSocket CDP endpoint for every session, so you can attach your own Playwright instance to a fully managed remote browser without installing or maintaining local Chromium infrastructure.