Key takeaways
- Let your coding agent control and inspect a live Chrome browser, instead of guessing how a page behaves.
- Get performance insights: the tool records traces with Chrome DevTools and extracts the findings.
- Debug with real data: network requests, screenshots and console messages are available to the agent.
- It works as an MCP server or as a command-line tool, and ships skills that teach agents to use it.
- The catches: browser content is exposed to the agent, usage statistics are on by default, and only Chrome is officially supported.
What it is
Chrome DevTools for agents is an open-source tool that lets a coding agent such as Antigravity, Claude, Cursor or Copilot control and inspect a live Chrome browser. It is released under the Apache-2.0 licence, and the code is on GitHub at ChromeDevTools/chrome-devtools-mcp.
The problem it solves
A coding agent that edits a web page normally cannot see how the page actually runs. Developers who use agents for front-end work end up copying console errors and performance numbers into the chat by hand.
This project connects the agent to a real Chrome instance, so it can look at the page itself.
What you get
- Performance insights: it uses Chrome DevTools to record traces and extract actionable findings.
- Debugging data: the agent can analyse network requests and take screenshots.
- Steadier automation: actions are automated in Chrome, and it automatically waits for action results.
- Two ways in: it works as an MCP server, and the package also includes a command-line tool.
- Skills for agents: the package includes skills that teach an agent how to use Chrome DevTools, and a slim option exposes a reduced set of tools for basic tasks.
The catches
- Exposure: the server lets the agent inspect, debug and modify any data in the browser or DevTools, so keep sensitive sessions out of it.
- Privacy: Google collects usage statistics by default, such as tool success rates, latency and environment information, and the README documents how to opt out.
- Performance tools may send trace URLs to the Google CrUX API to fetch real-user data, which can also be switched off.
- Update checks: by default the server periodically checks the npm registry for updates.
- Platform: other Chromium-based browsers may work, but this is not guaranteed.
- Upkeep: outdated skills can tell an agent to use tools or parameters that changed.
What we did not test
We did not install the tool, connect an agent or record a trace. Every claim here comes from a literal quote in the project's README or the GitHub API, and the feature descriptions are the project's own.
Verdict
Editors' view: this is aimed at developers who already use a coding agent on web projects and want it to work from real browser data. If you do not use agents, it adds little. If you try it, do so with a clean browser profile and read the privacy options first.
Sources
Need help adopting this, or have a software project in mind?
Tell us what you are building. The initial consultation is free. Email [email protected] or follow @zolnaltd.
Start a conversation →