Google Chrome DevTools Unleashes AI Debugging with New Model Context Protocol (MCP) Server

The Google Chrome development team has released a public preview of the Chrome DevTools Model Context Protocol (MCP) server for developers, opening DevTools to AI coding agents. The MCP server is designed so AI agents can directly debug web pages within Chrome, using DevTools’ debugging utilities, performance insights and diagnostic features.

For example, the server provides a tool called performance_start_trace, which can launch a site and record a performance trace for the AI agent to analyse and propose optimisations. Potential use cases include validating code changes, diagnosing network and console errors, simulating user behaviour, debugging live style and layout issues, and automating performance audits such as addressing excessive LCP values.

Developers interested in exploring its usage can consult the official Chrome DevTools MCP server documentation: according to Chrome DevTools MCP on GitHub.