Chrome DevTools MCP is a Model Context Protocol integration that lets AI tools see what the browser can see. It grants an agent direct access to inspect the DOM and to obtain rich performance traces, console logs and network traces — described in My LLM coding workflow going into 2026 as giving an agent “eyes”. Its source code is on GitHub at https://github.com/chromeDevTools/chrome-devtools-mcp.
Capabilities
Its purpose is to bridge the gap between static code analysis and live browser execution. By exposing runtime data from the browser to the agent, it removes the friction of manually switching context to relay that information, and allows UI testing to be automated directly through the LLM, so that bugs can be diagnosed and fixed based on actual runtime data.
Adoption & Ecosystem
Addy Osmani, whose earlier team built it, describes using it as part of his debugging and quality loop when coding with AI agents.