browse v1.4.0 — Performance Audit System
Your AI agent can now run a full web performance audit from the CLI.
perf-audit — one command, full report:
→ Core Web Vitals (LCP, CLS, TBT, FCP, TTFB, INP) with Google's thresholds
→ LCP critical path — traces the entire blocking chain from TTFB to render
→ Layout shift attribution — font swaps, missing image dimensions, ad injection
→ JS/CSS coverage — per-file used vs unused bytes → Image audit — format, lazy-load, fetchpriority, oversized detection
→ Font audit — font-display, preload, FOIT/FOUT risk
→ DOM complexity, render-blocking resources, long task attribution
detect — tech stack fingerprint:
→ 108 framework signatures across 12 categories
→ 55 SaaS platforms — Shopify app enumeration with per-app sizing, WordPress plugin list, Magento RequireJS waterfall
→ Third-party inventory with 88 known domain classifications
→ All detections in a single page.evaluate() call, under 200ms
Platform-aware recommendations.
Detects your stack and gives you specific fixes — not generic Lighthouse advice.
Shopify? It tells you which apps to remove.
WordPress? Which plugins are killing your TBT. Magento? Which RequireJS modules to bundle.
Built for agents.
Works with --mcp. Pipe it into Claude Code, Cursor, or any MCP client and let your AI fix your performance.