A page speed test that tells you what to do.
Core Web Vitals, Lighthouse scores and the list of fixes, ordered by how much time each one saves. Straight from Google's API, no account needed.
- Mobile, like Google ranks
- All four Lighthouse categories
- Results gone in 24 hours
What this measures
Google loads your page in a real browser on a simulated slow mobile connection, and reports four scores out of a hundred: performance, accessibility, best practices and SEO.
Behind the performance score sit the metrics that matter most:
Largest Contentful Paint — when the biggest thing on screen finished drawing. This is the one visitors feel.
Cumulative Layout Shift — how much the page jumped around while loading. The reason you tapped the wrong button.
Total Blocking Time — how long the page ignored taps because JavaScript was busy.
Lab data and field data are different things
The scores above come from one simulated load, which is what makes them repeatable enough to test a fix against. Separately, where Google has enough real Chrome traffic for your site, we show what actual visitors experienced over the last 28 days.
That second set is what Google ranks on, and it lags a fix by weeks. So the lab numbers tell you what to change; the field numbers tell you whether it worked, later.
Reading the numbers
Lighthouse's own bands. 90 is the line, which is why 88 stings.
- 90–100Good
- 50–89Needs work
- 0–49Poor
Page speed questions
Why is my score lower here than in Chrome DevTools?
DevTools measures your machine on your connection. This measures a simulated mid-range phone on a throttled connection, which is much closer to how most people will actually load your site — and much closer to what Google records.
It says my site is fine, but it feels slow.
Then trust the feeling and look at the field data section, if your site has enough traffic for Google to have any. A good lab score with poor field numbers usually means something that only happens on real devices: a heavy third-party script, or a slow server under load.
How often can I run it?
As often as you sensibly need to. There is a rate limit to stop abuse, and checking the same URL twice within fifteen minutes gives you the earlier result rather than spending a new measurement on it.