Free website speed test

Type the address in the box above and press Test. The tool asks Google's PageSpeed Insights for a mobile and a desktop run and shows both, with the real-visitor figures beside the lab ones. There is no sign-up and no email; ten tests a day per person.

Google's Core Web Vitals thresholds, which are what it ranks on: largest contentful paint under 2.5 seconds (good), 2.5 to 4 (needs improvement), over 4 (poor); interaction to next paint under 200 milliseconds; cumulative layout shift under 0.1. Those are the field numbers on this page, measured on real phones, not the lab score.

Because the mobile lab run simulates a mid-range phone on a slow 4G connection, which is closer to how most shoppers actually arrive. Oversized images, app scripts and web fonts that a desktop hides cost seconds on a phone. Fix mobile first; desktop usually follows.

Yes, and the effect is on the visitors you never see: the ones who leave before the page paints. We run online stores ourselves, and every speed fix we have made on our own stores showed up first as fewer bounces on mobile product pages. Google also uses the field Core Web Vitals as a ranking signal, so a slow page is both a conversion problem and a visibility one.

Google's. The tool calls the PageSpeed Insights API, which runs Lighthouse on Google's own machines — the same engine behind the score in Search Console and behind every speed report a consultant has ever emailed you. We show it unedited, including the parts that make our own pages look average.

The lab numbers are one visit by a machine on a simulated slow connection: repeatable, and not a real person. The field numbers are Chrome users who actually visited your site in the last 28 days, and they are the ones Google uses for ranking. A lab score of 95 with a 4-second field LCP means your real visitors are having a worse time than the test suggests, and that gap is the single most useful thing on this page.

Google only publishes it once a site has enough real traffic in its dataset. No field data is not a fault; it means too few Chrome users with reporting on have visited recently. The lab numbers still apply.

No. The request comes from Google's machines, not ours, and it is one page view. We cache the answer for 6 hours, so asking twice costs your server nothing.

Because each one asks Google's API and somebody else's server for work. Ten is plenty to test a site and its competitors; more than that is a script, and a script should be measuring its own sites with its own API key.

Whatever the opportunities list puts at the top — it is sorted by the time Google thinks you would save. In our experience on real stores the order is almost always: images that are bigger than the space they appear in, third-party scripts from apps nobody uses any more, and fonts loaded from somewhere else. The first two are usually most of the problem.