chrome-devtools-mcp
Chrome DevTools MCP, from the Chrome DevTools team, lets a coding agent such as Claude, Cursor, Copilot or Antigravity control and inspect a live Chrome browser. It exposes the power of DevTools to the agent: recording performance traces and extracting actionable insights, analysing network requests, reading console messages with source-mapped stack traces, and taking screenshots. Actions are automated with Puppeteer, which waits for results automatically.
It is aimed at front-end and full-stack developers who want their coding agent to verify its own changes in the browser: reproduce a bug, check why a request fails, read the console error behind a broken page, or investigate a slow load. A `--slim` mode covers basic browser tasks with fewer tools, and a CLI is available for use without MCP.
Getting started: requires a current Node.js LTS, npm and a stable Chrome. Add `npx -y chrome-devtools-mcp@latest` to your MCP client configuration, then ask the agent to check a page, for example its performance. Pass `--no-usage-statistics` if you do not want usage data sent to Google.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-10-23
- Commits, last 90 days
- 100+
- Releases, last 90 days
- 7
- Latest release
- chrome-devtools-mcp-v1.12.0 (2026-10-22)
- Open issues
- 106
- Recent issues with at least one comment
- 3/4
Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #268 A tool to "Select an element in the page to inspect it"26 comments · 2025-10-04
- #1921 Browser hangs / crashes when MCP connects to Chrome with many tabs18 comments · 2026-04-21
- #2584 Logpoints: log runtime values without modifying source code13 comments · 2026-08-17
- #3070 Allow multiple sessions / processes / instances of Chrome10 comments · 2026-10-11
- #889 The DevTools MCP is hardcoded to look for Chrome in standard locations9 comments · 2026-02-05
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Who it fits
- You want an agent to record a Chrome performance trace and get insights from it; the README lists performance_start_trace, performance_stop_trace and performance_analyze_insight.
- You need to inspect network requests, console messages and screenshots of a page while debugging it.
- You want the agent to attach to a Chrome you already have running, using --browser-url, --ws-endpoint or --autoConnect (Chrome 144+).
- You need Firefox or Safari: the project is built on Chrome DevTools and targets Chrome.
- You want a general-purpose automation or test runner across browsers; Playwright MCP is the closer fit.
- You cannot run Node.js 20.19 or newer, which the README lists as a requirement.
How it differs
vs playwright-mcp★ 38.7kChrome DevTools MCP is maintained under the ChromeDevTools organisation and is strongest at debugging and performance analysis in Chrome. Playwright MCP covers more browsers and test-style automation. They can be installed side by side, as each is a separate MCP server.
vs browser-tools-mcp★ 7.3kBrowser Tools MCP also offers Lighthouse audits and console and network logs, but it needs a Chrome extension plus a local Node server in addition to the MCP server. Chrome DevTools MCP launches or attaches to Chrome directly and needs no extension.
Compatibility
Quick start
Use cases
Alternatives
Featured in these lists
Comments
No comments yet. Be the first!