Enter any public URL and this tool asks Google’s free PageSpeed Insights service to test it live — straight from your browser, no sign-up and no API key. You get the Lighthouse lab score and metrics, plus real-user field data (CrUX) when Google has enough of it for the page.
A check usually takes 10–40 seconds — Google runs a real Lighthouse test on its own servers. Press Enter in the URL field to start.
Asking Google to test the page…
Elapsed 0.0s · please keep this tab open.
The check could not be completed
This tool uses Google’s free, shared PageSpeed Insights quota, so a failure is often temporary — retrying in a few minutes usually works.
Scored 0–100 from a single simulated load (lab data). 90–100 is good, 50–89 needs work, below 50 is poor. Lab scores move between runs — judge by the metrics below, not one number.
Measured by Lighthouse in a controlled run. Lab data has no true INP, so Total Blocking Time is shown as the responsiveness proxy — labelled honestly as such.
How long the biggest text or image block takes to appear — the moment the page feels loaded. Good is 2.5 seconds or under; over 4 seconds is poor. Slow servers, huge hero images and render-blocking files are the usual causes.
How much the page jumps around while loading — buttons moving as ads, fonts or images arrive. It is a unitless score: 0.1 or under is good, over 0.25 is poor. Reserve space for images and embeds to fix it.
How quickly the page visibly responds after a tap, click or key press. Only real-user field data can measure it — a lab run has no genuine interactions, which is why the lab section shows Total Blocking Time instead. Good is 200 ms or under; over 500 ms is poor.
The total time the main thread was blocked by long JavaScript tasks between first paint and interactivity. It is the lab stand-in for INP: 200 ms or under is good, over 600 ms is poor. Heavy scripts and third-party widgets drive it up.
FCP is when the first text or image paints (good: 1.8 s or under). Speed Index measures how quickly content is visibly populated on average (good: 3.4 s or under). Time to Interactive is when the page becomes fully usable (good: 3.8 s or under).
Lab data is one controlled test — repeatable and great for debugging, but not what visitors actually experienced. Field data is the 75th percentile of real Chrome visits over 28 days, and it is the field numbers Google uses for the Core Web Vitals assessment. New or low-traffic pages often have no field data yet.
Live Core Web Vitals from Google PageSpeed Insights — lab metrics, field data and plain-English thresholds.
100% free · No sign-up · No daily limits · No watermarks · Processed in your browser — files never leave your device
Run a live Core Web Vitals check on any public URL using Google’s free PageSpeed Insights service: a Lighthouse performance score, the lab metrics (LCP, CLS, Total Blocking Time, FCP, Speed Index, TTI) and real-user field data when Google has it — no API key, no sign-up.
A shop owner tests https://asadwebs.com/ on Mobile and gets a score of 92 with LCP at 1.8 s and CLS at 0.04 — both Good — but Total Blocking Time over 600 ms would flag Poor, pointing straight at heavy JavaScript as the thing to fix before touching images or fonts. The field section then shows how real Chrome visitors did over the last 28 days, which can tell a different story from the single lab run.
The checker sends the URL from your browser to Google’s PageSpeed Insights API (v5, keyless, performance category only) and reads two parts of the response. Lab data comes from lighthouseResult: the performance category score and the audits for Largest Contentful Paint, Cumulative Layout Shift, Total Blocking Time, First Contentful Paint, Speed Index and Time to Interactive, plus the experimental Interaction to Next Paint audit on the runs where Google includes it. Field data comes from loadingExperience (falling back to originLoadingExperience for the whole site): the CrUX 75th percentiles for LCP, CLS and INP and their FAST / AVERAGE / SLOW categories. Every chip uses Google’s published thresholds — LCP good at 2.5 s or under, CLS at 0.1 or under, TBT and INP at 200 ms or under.
Three measures of real page experience that Google uses as a ranking signal: Largest Contentful Paint (how fast the main content appears, good at 2.5 s or under), Interaction to Next Paint (how fast the page responds to taps and clicks, good at 200 ms or under) and Cumulative Layout Shift (how much the layout jumps around, good at 0.1 or under). A page passes when all three are good at the 75th percentile of real visits.
INP needs a real person interacting with the page, and a lab test has no real interactions to measure. Total Blocking Time — how long JavaScript blocks the main thread — is the accepted lab stand-in: pages with low TBT almost always respond quickly in the field. When Google’s response includes its experimental lab INP audit, this tool shows that too, clearly labelled as a lab estimate.
Field data comes from the Chrome User Experience Report, which only publishes a page once enough Chrome users have visited it over a rolling 28-day window. New, low-traffic or recently-moved pages show the honest “not enough real-user data” state instead. If the page has no data but its site does, the tool shows the site-wide (origin) numbers and labels them as such.
Every run is a fresh simulated load: server response, network conditions on Google’s tester and third-party scripts all vary a little, so lab scores commonly move a few points between runs. Treat a single run as a snapshot, fix what the metrics flag, and look at field data for the stable picture.
Mobile first: most web traffic is mobile, the mobile test uses a slower simulated phone and network, and Google indexes mobile-first. A page can score 95 on Desktop and 60 on Mobile — the mobile number is the one your visitors feel.
The tool calls Google’s free keyless quota, which is shared and can run out at busy times (HTTP 429). Nothing is wrong with your URL — wait a few minutes and run the check again.
What is missing, broken, or could be better on this tool? Tell us — we read every suggestion.
✅ Thanks — your suggestion has been received. Suggestions like yours shape what we build next.