lighthouse
MCP server that enables AI agents to perform comprehensive web audits using Google Lighthouse with 13+ tools for performance, accessibility, SEO, and security analysis.
claude mcp add --transport stdio danielsogl-lighthouse-mcp-server npx @danielsogl/lighthouse-mcp@latest
How to use
The Lighthouse MCP Server provides a dedicated MCP interface to run Google Lighthouse audits via a server you can query with your MCP client. It enables LLMs and AI agents to perform comprehensive web performance analyses, including core web vitals, accessibility checks, SEO insights, and security assessments. Through the MCP protocol, you can initiate audits, fetch results, and tailor Lighthouse runs with CLI-like options such as disabling or enabling headless mode, passing Chrome flags, and connecting to a persistent Chrome profile for authenticated sessions. The server is designed to integrate with MCP clients (VS Code, Cursor, Windsurf, Claude Desktop, etc.) so you can orchestrate automated audits as part of larger workflows.
How to install
Prerequisites: Node.js 22.0.0 or newer, a compatible MCP client, and a network environment that can reach npm registry.
- Install Node.js (if not already installed).
- Ensure you have npm or pnpm configured to install npm packages.
- Start the Lighthouse MCP server via your MCP client or command line:
- Using npx (as shown in the README): npx @danielsogl/lighthouse-mcp@latest
-
(Optional) Pin to a specific version: npx @danielsogl/lighthouse-mcp@1.2.3
-
Verify the MCP server appears in your MCP client and is ready to accept commands.
-
If you plan to persist sessions, prepare a Chrome profile path and use the appropriate flags when invoking the server via the MCP client.
Additional notes
Tips and common considerations:
- The server relies on Lighthouse (and a Chrome/Chromium instance managed by Lighthouse) to perform audits; ensure Chrome/Chromium is accessible in your environment.
- When using persistent sessions, profile-path and user-data-dir options help maintain cookies and local storage between audits.
- You can pass additional Chrome flags via --chrome-flag to customize the browser environment; for some flags, prefer the trailing form like --chrome-flag=--disable-gpu.
- Use the MCP client’s standard mcpServers configuration to integrate with other tools; the example in README demonstrates the canonical Lighthouse setup.
- If you encounter network or permission issues, verify that your environment permits executing npx and fetching the Lighthouse MCP package from npm.
Related MCP Servers
context7
Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors
obsidian -tools
Add Obsidian integrations like semantic search and custom Templater prompts to Claude or any MCP client.
MiniMax -JS
Official MiniMax Model Context Protocol (MCP) JavaScript implementation that provides seamless integration with MiniMax's powerful AI capabilities including image generation, video generation, text-to-speech, and voice cloning APIs.
mcp-bundler
Is the MCP configuration too complicated? You can easily share your own simplified setup!
akyn-sdk
Turn any data source into an MCP server in 5 minutes. Build AI-agents-ready knowledge bases.
promptboard
The Shared Whiteboard for Your AI Agents via MCP. Paste screenshots, mark them up, and share with AI.