Free SEO tool · Speed

Website speed test & Core Web Vitals

See in under a minute how fast your page loads on mobile and desktop: Lighthouse scores, real-user data from Chrome (CrUX) and what is worth fixing first.

4 Lighthouse scoresCrUX real-user data from Google$0 no account needed
Lab · Google PageSpeed Insights Lighthouse + CrUX
Device

Start a test whenever you're ready

Enter a URL, pick mobile or desktop, and hit “Run test.” The analysis runs on Google's servers and usually takes 15–40 seconds.

  • 4 Lighthouse scores
  • Real-user Core Web Vitals
  • Page screenshot
  • What to fix first

Privacy: the URL you test is sent to Google PageSpeed Insights, which loads the page like any other visitor — exactly what happens when you use pagespeed.web.dev. We don't receive or store the URL; your last 5 tests are saved only in your own browser.

How to use the speed test

Enter the full URL of the page you want to check. Don't stop at the homepage: test a service page, a product page and a blog post too, because each template ships different images, scripts and, often, different problems. Choose Mobile for the view closest to how Google evaluates you (Google indexes mobile-first), then repeat on Desktop if most of your customers browse from a computer.

The test runs on Google's servers with a simulated phone CPU and network. Results usually appear within 15–40 seconds, and your last five runs stay saved in the browser so you can compare before and after a change.

What the results mean

  • Lighthouse scores (0–100) are lab measurements. The performance score is a weighted blend of five metrics: Total Blocking Time (30%), Largest Contentful Paint (25%), Cumulative Layout Shift (25%), First Contentful Paint (10%) and Speed Index (10%). 90 and above is green, 50–89 is orange, below 50 is red.
  • Real-user data (CrUX) comes from Chrome users who visited the page over the previous 28 days. What counts is the 75th percentile: if three out of four visits fall within the “good” threshold, the metric passes. The colored bar shows how visits split between good, needs improvement and poor.
  • Opportunities are sorted by the estimated time you could save. They are lab estimates, useful for prioritizing, not a promise of a specific score.

Core Web Vitals thresholds

MetricGoodNeeds workPoor
LCP — main content loads≤ 2.5 s2.5–4 s> 4 s
INP — response to taps≤ 200 ms200–500 ms> 500 ms
CLS — layout stability≤ 0.10.1–0.25> 0.25

For the full explanation with examples, read our guide to Core Web Vitals and site speed.

What to fix first

  1. The hero image. It is usually the LCP element. Serve it in a modern format (AVIF or WebP) at the size it is actually displayed, without lazy loading and with high fetch priority. Our guide to image SEO, WebP and alt text covers the details.
  2. Third-party scripts. Chat widgets, ad pixels and social embeds block the main thread and hurt INP. Defer them, or load them only on the pages that need them.
  3. Server and caching. A slow time to first byte (TTFB) delays everything after it. Page caching, compression and a CDN fix most of it.
  4. Layout stability. Reserve space for images, ads and the cookie banner so they don't push the text down while a visitor is reading.
  5. Fonts. Use font-display: swap, load only the weights you need and subset the files to the characters your audience actually reads.

Common mistakes

The most frequent one is chasing a score of 100 instead of real-user experience: a site can score 95 in the lab and still fail Core Web Vitals on visitors’ phones. The second is testing a single page, once. The third is stacking three or four overlapping “optimization” plugins that conflict and break features. Speed is one part of technical SEO, next to crawling, indexing, structure and structured data; if you are building a new site, it is far cheaper to plan for it from day one, as in an SEO-optimized website build.

If the report shows problems you don't know how to solve, an SEO audit connects speed to the other factors that influence your rankings — or get in touch and we'll go through the report with you.

Frequently asked questions

Speed test FAQ

Why does the score change from one test to the next?

A Lighthouse score comes from a single page load on a Google server with a simulated network and CPU. Your server's load, ads and third-party scripts differ from run to run, so a few points of variation is normal. For decisions, look at the trend across several tests and, above all, at the real-user CrUX data.

What matters to Google: the Lighthouse score or Core Web Vitals?

According to Google, its page experience signals draw on real Chrome user data (the Core Web Vitals field data in CrUX), not the lab score. Lighthouse is the diagnostic tool: it shows why a page is slow and what you can fix. More in our guide to Core Web Vitals.

Why is there no real-user (CrUX) data for my site?

Google publishes field data only for pages and origins with enough traffic from Chrome users who opted in to usage statistics. New or low-traffic sites often have none yet. In that case the tool falls back to origin-level data, and if that is missing too, you rely on the lab measurements.

Why does the test fail or say the limit was reached?

The test uses the public Google PageSpeed Insights API, which has a free request quota. If the quota is used up, try again later or run the test directly on pagespeed.web.dev — the link appears automatically in the error message. Pages that require a login, block bots or return errors can't be analyzed.

What numbers should a fast site have?

Google's “good” thresholds, measured at the 75th percentile of visitors, are an LCP of 2.5 seconds or less, an INP of 200 ms or less and a CLS of 0.1 or less. If your site is far from those, our technical SEO team can find and fix the causes.

Slow site? We'll find the cause and fix it.

We audit images, scripts, server and caching, fix what slows the page down, then track the results against real-visitor data.