Chrome DevTools Protocol

Chrome DevTools is a set of web authoring, debugging, and performance analysis tools built into Google Chrome and other Chromium-based browsers. Maintained by the Chromium/Chrome team at Google, it provides DOM inspection, JavaScript debugging, network monitoring, and performance profiling for web developers.

Chrome DevTools for agents (chrome-devtools-mcp) is an MCP server from the Chrome DevTools project that lets coding agents control and inspect a live Google Chrome or Chrome for Testing browser. It uses Puppeteer for browser automation, including navigation and actions that wait for results, and exposes DevTools capabilities for recording and inspecting performance traces and insights, analyzing network requests, taking screenshots, and inspecting browser console messages with source-mapped stack traces. A CLI is also provided for use without MCP, with configuration options including headless, isolated, slim, concurrent-session, persistent-profile, WebSocket, and Android-debugging modes. It requires Node.js LTS, a current stable Chrome version or newer, and npm. Browser contents are exposed to MCP clients; performance tools may query the Google CrUX API, and usage statistics and update checks are enabled by default with an opt-out flag.

Puppeteer is a JavaScript library that provides a high-level API for controlling Chrome or Firefox through the DevTools Protocol or WebDriver BiDi. It runs browsers headlessly by default and can automate navigation, viewport settings, keyboard input, locating and interacting with page elements, and browser shutdown. The library can be used to render web content into PDFs; the video describes substantial memory usage and slow cold starts as drawbacks in that use case. The `puppeteer` package downloads a compatible browser during installation, while `puppeteer-core` provides the library without downloading Chrome.

Browser Use Pi is a TypeScript web agent built on Pi Mono. It combines a persistent V8 REPL with raw Chrome DevTools Protocol control: the agent writes JavaScript, uses accessibility-tree data and screenshots, and builds browser helpers as needed. It supports persistent sessions, saved logins, workspaces, streaming, hooks, typed results, follow-up tasks, and cloud browsers or a user's own Chrome; runs can be limited by steps, time, or cost, with partial work retained. Interaction highlights, recordings, and GIF exports can show the agent's work. It is distributed as the @browser_use/pi package for Node 22.19+ or Bun 1.3.14+, and its quickstart uses Browser Use Cloud without requiring a local Chrome installation.