Skip to main content
Glama

Lighthouse MCP Server

通过 MCP 协议在 Claude 中直接运行 Lighthouse 网页性能审计,获取 Core Web Vitals、性能评分和优化建议。

功能

提供两个 MCP 工具:

audit_url — 单页审计

对一个 URL 运行 Lighthouse,返回:

  • Core Web Vitals: LCP、CLS、INP

  • 其他性能指标: FCP、TTFB、Speed Index、TBT

  • 分类评分: Performance / Accessibility / SEO / Best Practices (0-100)

  • 优化建议: 按预计可节省时间排序的改进机会

audit_batch — 批量审计

对多个 URL(最多 10 个)并行运行 Lighthouse,返回每个页面的核心指标,适合跨页面对比。

Related MCP server: LightScout MCP

前置条件

  • Node.js >= 18

  • Google Chrome — Lighthouse 依赖 Chrome 运行,chrome-launcher 会自动查找。macOS 上安装 Chrome 即可。

安装

git clone <repo-url>/lighthouse-mcp.git
cd lighthouse-mcp
npm install
npm run build

在 Claude Code 中使用

在 ~/.claude/claude_desktop_config.json(或对应配置文件)中添加:

{
  "mcpServers": {
    "lighthouse": {
      "command": "node",
      "args": ["/path/to/lighthouse-mcp/build/index.js"]
    }
  }
}

配置后重启 Claude Code,即可使用:

帮我审计一下 https://example.com 的性能
对比这三个页面的性能: https://a.com https://b.com https://c.com
用桌面模式审计 https://example.com

输出示例

{
  "url": "https://example.com",
  "device": "mobile",
  "scores": { "performance": 85 },
  "coreWebVitals": {
    "lcp_ms": 2500,
    "cls": 0.1,
    "inp_ms": 150
  },
  "otherMetrics": {
    "fcp_ms": 1800,
    "ttfb_ms": 200,
    "speedIndex_ms": 3000,
    "tbt_ms": 200
  },
  "opportunities": [
    { "title": "Reduce unused JavaScript", "potentialSavingsMs": 500 }
  ]
}

技术栈

  • @modelcontextprotocol/sdk — MCP 协议实现

  • lighthouse — Google Chrome 网页审计工具

  • chrome-launcher — 无头 Chrome 启动

  • TypeScript + ESM

License

MIT

Available Tools

2 tools
audit_batchA

Run Lighthouse performance audits on multiple URLs in parallel. Returns a summary comparing Core Web Vitals across all URLs, plus detailed metrics for each.

Inputs: urls (required), device.

audit_urlA

Run a Lighthouse performance audit on a URL. Returns Core Web Vitals (LCP, CLS, INP), other performance metrics (FCP, TTFB, Speed Index, TBT), category scores (0-100), and a prioritized list of improvement opportunities with estimated savings.

Inputs: url (required), device, categories.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.1/5.0

    Scored across 2 tools

    Disambiguation5/5

    The two tools are clearly distinct: audit_url handles a single URL while audit_batch handles multiple URLs. There is no overlap or ambiguity in their purposes.

    Naming Consistency5/5

    Both tools use the consistent verb_noun pattern (audit_url, audit_batch), making the naming predictable and easy to follow.

    Tool Count3/5

    With only 2 tools, the server feels thin for a full-featured Lighthouse server, though the tools cover the core audit functionality. It is below the typical 3-15 range.

    Completeness4/5

    The tools cover both single and batch audits, returning comprehensive metrics and opportunities. Minor gaps include lack of custom configuration options (e.g., device, categories) or historical comparison, but these are not critical.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      B
      maintenance
      MCP server that enables AI agents to perform comprehensive web audits using Google Lighthouse with 13+ tools for performance, accessibility, SEO, and security analysis.
      11
      1,975 npm
      75
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      Core Web Vitals analysis powered by Lighthouse. Four tools: analyze a URL, compare two URLs, check against thresholds, or crawl an entire site. Works with Claude Code, Cursor, Windsurf, and any MCP-compatible AI tool.
      4
      64 npm
      MIT