Website Performance Analyzer

Audit any URL for Core Web Vitals, SEO, accessibility, and page weight — then get a prioritized, human-readable list of exactly what to fix. Unlike tools that dump raw Lighthouse JSON, this actor produces structured reports with pass/fail ratings and actionable recommendations sorted by impact.

What It Measures

Core Web Vitals (with pass/fail ratings)

• LCP (Largest Contentful Paint) — how fast the main content loads. Good: < 2.5s
• FCP (First Contentful Paint) — when the first pixel of content appears. Good: < 1.8s
• CLS (Cumulative Layout Shift) — how much the layout shifts during load. Good: < 0.1
• TTFB (Time to First Byte) — server response speed. Good: < 800ms

Each metric receives one of three ratings: good, needs-improvement, or poor — matching Google's CWV thresholds.

Page Stats

Total transfer size (with human-readable formatting), resource count, per-type breakdown (scripts, stylesheets, images, fonts, etc.), and a count of render-blocking resources that delay page rendering.

SEO Basics

Title tag (presence + length 10–70 chars), meta description (presence + length 50–160 chars), viewport meta tag, canonical URL, and H1 heading (presence + count).

Accessibility Basics

Images without alt text, links without descriptive text or aria-label, form inputs without associated labels, and a heuristic contrast check across visible text elements.

Performance Scores

Each URL receives three scores (0–100): Performance, SEO, and Accessibility — calculated directly from the measured metrics using the same thresholds as industry-standard auditors.

Prioritized Recommendations

Every result includes a ranked list of fixes sorted high → medium → low priority. Each recommendation includes:

• Category (performance, seo, or accessibility)
• A plain-English title
• A description explaining what to fix and why, with specific values from the page

Examples: "LCP is 4,200ms — add <link rel=preload> for your hero image", "3 images missing alt text — add descriptive alt attributes", "No meta description found — add 50–160 character description".

Input Options

Field	Default	Description
urls	[{"url": "https://crawlee.dev"}]	URLs to analyze. Each entry: { "url": "...", "label": "..." }
device	mobile	Viewport mode: mobile (375×812) or desktop (1280×800)
waitForNetworkIdle	true	Wait for network activity to settle before collecting metrics
maxConcurrency	3	URLs analyzed in parallel (1–10)
requestTimeout	60000	Page load timeout in milliseconds (10,000–120,000)

All fields have defaults — the actor runs zero-config.

Output Format

Each URL produces one dataset record with:

• coreWebVitals — LCP, FCP, CLS, TTFB with values, ratings, and units
• pageStats — total requests, transfer size, resource breakdown by type
• seo — title, meta description, viewport, canonical, H1 checks
• accessibility — image alt, link text, input label, contrast issue counts
• scores — performance (0–100), seo (0–100), accessibility (0–100)
• recommendations — prioritized fix list with title + description
• analyzedAt — ISO 8601 timestamp

Competitive Advantage

Most performance tools produce raw JSON that requires interpretation. This actor produces decision-ready output: every result tells you what's wrong and what to do about it, ranked by impact. Useful for developers auditing their own sites, agencies producing reports for clients, and SEO/performance monitoring workflows.

Pricing

This actor charges per URL analyzed via Apify's Pay Per Event billing. You only pay for URLs that complete analysis — failed navigations are not charged.

---

FAQ

Can I analyze multiple URLs in one run? Yes — pass any number of URLs in the urls array. They are processed in parallel up to maxConcurrency.

How is the performance score calculated? LCP (40%), FCP (25%), CLS (25%), TTFB (10%) — each metric earns 100 points for "good", 60 for "needs-improvement", 20 for "poor". Scores are weighted and summed.

Does this run real Lighthouse? No — it uses Playwright's browser Performance APIs (Navigation Timing, Paint Timing, PerformanceObserver) which capture the same underlying metrics. Results are comparable to Lighthouse field data.

What devices are supported? Mobile (iPhone-style, 375×812, 2× DPR) and Desktop (1280×800, 1× DPR). Mobile is the default since Google uses mobile-first indexing.

Can I analyze pages behind a login? Not directly in the current version. Authentication flows require additional configuration.

How accurate is the contrast check? The contrast check is heuristic — it samples up to 100 text elements and calculates WCAG contrast ratios from computed styles. It may miss some issues and flag some false positives on elements with complex backgrounds. Use a dedicated accessibility scanner for comprehensive WCAG auditing.