[BA-001] What is CDP, Chrome's DevTools Protocol for remote browser control

If you ever used Puppeteer, Playwright or even Selenium WebDriver, you were talking to the browser through CDP without knowing it.

CDP stands for Chrome DevTools Protocol. It is a protocol that allows you to inspect, debug and control a Chromium based browser programmatically. It works over WebSocket and exposes almost everything the browser can do.

To test it yourself, first install the websocket-client library:

pip install websocket-client

Then close all Chrome instances and start Chrome with remote debugging enabled:

chrome.exe --remote-debugging-port=9222

Open another terminal and run this script:

import json
import requests
import websocket

# Get the WebSocket URL from Chrome
resp = requests.get("http://localhost:9222/json/version")
ws_url = resp.json()["webSocketDebuggerUrl"]

ws = websocket.create_connection(ws_url)

# Open a new tab and navigate to a page
cmd = {
    "id": 1,
    "method": "Target.createTarget",
    "params": {"url": "https://example.com"}
}
ws.send(json.dumps(cmd))
result = json.loads(ws.recv())
print("Tab opened:", result["result"]["targetId"])

ws.close()

If you run this, Chrome will open a new tab pointing to example.com. The script prints the targetId of the new tab.

You can also take a screenshot of a page without any screenshot tool. After opening the tab, send this command:

cmd = {
    "id": 2,
    "method": "Page.captureScreenshot",
    "params": {"format": "png"}
}
ws.send(json.dumps(cmd))
result = json.loads(ws.recv())

with open("screenshot.png", "wb") as f:
    import base64
    f.write(base64.b64decode(result["result"]["data"]))

print("Screenshot saved as screenshot.png")

Run the full script below. It will open a new tab, navigate to example.com, take a screenshot and save it to disk.

import json
import base64
import requests
import websocket

resp = requests.get("http://localhost:9222/json/version")
ws_url = resp.json()["webSocketDebuggerUrl"]

ws = websocket.create_connection(ws_url)

cmd_open = {
    "id": 1,
    "method": "Target.createTarget",
    "params": {"url": "https://example.com"}
}
ws.send(json.dumps(cmd_open))
ws.recv()

cmd_ss = {
    "id": 2,
    "method": "Page.captureScreenshot",
    "params": {"format": "png"}
}
ws.send(json.dumps(cmd_ss))
result = json.loads(ws.recv())

with open("screenshot.png", "wb") as f:
    f.write(base64.b64decode(result["result"]["data"]))

print("Screenshot saved as screenshot.png")
ws.close()

CDP is the low level protocol that tools like Puppeteer wrap into a friendlier API. Understanding it helps when you need to do something those tools do not expose.

That's all for now.
Thanks for reading!