PageVitals

@PageVitals show me the Core Web Vitals scores for our homepage and product pages over the last 7 days

Based on PageVitals monitoring data, here's a summary of Core Web Vitals performance: Homepage (example.com): - LCP: 1.8s (Good) — stable across the week - FID: 45ms (Good) — slight improvement on mobile - CLS: 0.08 (Needs Improvement) — spike on Dec 18th Product Pages (/products/*): - LCP: 2.4s (Needs Improvement) — desktop performs better - FID: 120ms (Needs Improvement) — consistent issue - CLS: 0.05 (Good) The homepage CLS spike correlates with a banner deployment. Product pages show consistent interaction delay that may warrant investigation.

This read-heavy example shows PageVitals' strength in historical performance tracking. The MCP requires API key authentication and returns data only for websites already configured in your account. If pages aren't being monitored yet, you'll need to add them first using the Add Configured Pages tool.

@PageVitals create a performance budget for our checkout flow: LCP under 2.5s, FID under 100ms, and CLS under 0.1

Performance budget created successfully for the checkout flow. Budget ID: bgt_9x4k2m Website: example.com Thresholds configured: - Largest Contentful Paint: ≤ 2.5s - First Input Delay: ≤ 100ms - Cumulative Layout Shift: ≤ 0.1 This budget is now active. PageVitals will flag any synthetic test runs or real-user measurements that exceed these thresholds. Depending on your account alerting settings, the website owner may receive notifications when budget violations occur.

This write example demonstrates persistent configuration changes. Creating a budget affects future monitoring behavior and may trigger alerts to your team. The operation consumes account resources and changes how PageVitals evaluates performance, so it should only be invoked when the user explicitly provides threshold values and confirms the action.

@PageVitals analyze our slowest pages from the last month and draft recommendations for the three biggest performance bottlenecks

After reviewing PageVitals data across all monitored pages, here are the three critical bottlenecks: 1. **Third-party script blocking (affects 12 pages)** Problem: Analytics and chat widgets delay FID by 180-240ms on mobile. Recommendation: Defer non-critical scripts using async/defer attributes, or load them after user interaction. 2. **Unoptimized hero images (affects 8 pages)** Problem: Above-the-fold images average 1.2MB, pushing LCP to 3.1s. Recommendation: Implement responsive images with WebP format and lazy-loading for below-the-fold content. 3. **Layout shift from dynamic content (/blog/*)** Problem: CLS averages 0.18 due to ad slots loading after initial render. Recommendation: Reserve explicit space for ad containers with min-height CSS to prevent reflow. These three fixes would improve ~65% of your monitored pages.

This synthesis example pairs PageVitals' monitoring data with AI reasoning to surface actionable insights. The MCP provides raw metrics; the AI interprets patterns and drafts recommendations. Results depend on having sufficient historical data in your PageVitals account — newly added websites may not have enough signal for meaningful analysis.

When PageVitals makes sense for a product launch sprint

A 6-person team shipping a redesigned marketing site in two weeks needs to catch regressions before customers do. PageVitals wins here because you can add the staging URL, create budgets for Core Web Vitals thresholds, and run multistep tests (checkout flow, signup funnel) without writing Lighthouse scripts. The 28-tool surface means you'll configure once and mostly read metrics after that—perfect for a launch sprint where you check dashboards daily but don't want to babysit CI. If your deploy cadence is slower than weekly, though, you're paying for monitoring you won't use between releases. Best fit: teams that ship user-facing changes at least twice a month and need a single place to track performance drift across 5-15 critical pages.

PageVitals as the source-of-truth for agency deliverables

A 3-person dev shop managing 12 client sites needs to prove they're hitting contractual performance targets (LCP under 2.5s, CLS under 0.1). PageVitals solves this because each client gets a separate website entry, you set budgets that match the SLA, and the API lets you pull monthly reports without manual screenshots. The persistent monitoring means you catch when a client's content team adds an unoptimized hero image—before the client notices. The trade-off: if you're only checking performance at handoff (not ongoing), you don't need 24/7 synthetic tests; a one-time Lighthouse run is cheaper. This MCP pays off when you're contractually on the hook for uptime and performance over months, not just at launch.

When multistep tests catch revenue-killing slowdowns

A 10-person e-commerce team sees conversion drop 8% after a Friday deploy and needs to know if checkout performance regressed. PageVitals' multistep test tool lets you script the add-to-cart → payment → confirmation flow and run it on mobile and desktop profiles, surfacing where the funnel got slow. This is faster than writing Puppeteer scripts and more focused than full-page Lighthouse audits. The limitation: if your checkout has complex auth or requires live payment tokens, the synthetic test setup gets brittle—you'll spend more time maintaining test credentials than analyzing results. Use this when your funnel is mostly public-facing (guest checkout, demo accounts) and you deploy checkout code at least weekly. Skip it if your checkout is behind SSO or you only ship monthly.