Chrome DevTools MCP is a Model Context Protocol (MCP) server that lets a coding agent such as Antigravity, Claude Code, Cursor, or Copilot control and inspect a live Chrome browser. It exposes Chrome DevTools capabilities as MCP tools for browser automation, debugging, and performance analysis, and a separate CLI offers the same capabilities without MCP.
Automation runs through Puppeteer, which drives Chrome and waits for action results automatically. The server starts the browser only when a client calls a tool that needs one, and it can also attach to an already running Chrome instance, run headless or isolated, handle concurrent sessions, or debug on Android. A slim mode exposes a smaller tool set for basic browser tasks.
The server exposes browser content to MCP clients, which can inspect and modify any data in the browser. It officially supports Google Chrome and Chrome for Testing. Performance tools may send trace URLs to the Chrome User Experience Report (CrUX) API to add real-user field data, and Google collects usage statistics by default, both of which can be turned off. Agent builders can also build an integrated browser subagent on top of it, as the Gemini CLI browser agent does.
Features
- Performance insights: records DevTools traces and extracts performance findings, optionally with CrUX field data alongside lab data
- Browser debugging: analyses network requests, takes screenshots, and reads console messages with source-mapped stack traces
- Automation: Puppeteer-driven actions in Chrome that wait for results before returning
- Slim mode: a reduced tool set for basic browser tasks
- Flexible browser setup: headless or isolated runs, persistent profiles, attaching to a running Chrome, WebSocket connections, and Android debugging
- CLI: the same DevTools capabilities without an MCP client
- Browser subagent base: a foundation for building an integrated browser subagent into agent products
