PageSpeed MCP Server

A Model Context Protocol (MCP) server that extends AI assistant capabilities with PageSpeed Insights functionality. This server acts as a bridge between AI models and Google’s PageSpeed Insights API, enabling detailed performance analysis of websites.

Overview

The PageSpeed MCP server is designed to enhance AI assistants’ capabilities by allowing them to perform comprehensive web performance analysis. When integrated, AI models can request and interpret detailed performance metrics, Core Web Vitals, and other critical web performance data for any given URL.

Installation

Installing via Smithery

To install PageSpeed Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install mcp-pagespeed-server --client claude

Manual Installation

npm install pagespeed-mcp-server

Configuration

Add the PageSpeed MCP to your AI assistant’s(claude in this case) configuration file:

{
    "pagespeed": {
        "command": "node",
        "args": ["path/to/mcp-pagespeed-server/dist/index.js"],
        "env": {
           "GOOGLE_API_KEY": ""
        }
    }
}

Detailed Capabilities

Performance Metrics Analysis

• First Contentful Paint (FCP)
• Largest Contentful Paint (LCP)
• Time to Interactive (TTI)
• Total Blocking Time (TBT)
• Cumulative Layout Shift (CLS)
• Speed Index
• Time to First Byte (TTFB)

Best Practices Assessment

• HTTPS usage
• JavaScript error monitoring
• Browser console warnings
• Deprecated API usage
• Image aspect ratio analysis
• Link security checks

SEO Analysis

• Meta description validation
• Robots.txt validation
• Structured data validation
• Crawlable links verification
• Meta tags assessment
• Mobile friendliness

Accessibility Audits

• ARIA attribute validation
• Color contrast checking
• Heading hierarchy analysis
• Alt text verification
• Focus management assessment
• Keyboard navigation testing

Resource Optimization

• Image optimization suggestions
• JavaScript bundling analysis
• CSS optimization recommendations
• Cache policy validation
• Resource minification checks
• Render-blocking resource identification

API Response Structure

The MCP server provides detailed JSON responses including:

{
    "lighthouseResult": {
        "categories": {
            "performance": { /* Performance metrics */ },
            "accessibility": { /* Accessibility results */ },
            "best-practices": { /* Best practices audit */ },
            "seo": { /* SEO findings */ }
        },
        "audits": {