Submit a URL and the test calls Google PageSpeed Insights, which runs Lighthouse on the page and returns a performance score plus Core Web Vitals such as LCP and CLS.
Pro tip: Test a real content URL, not the homepage, to see speed problems most visitors actually hit.
This tool is a front end for Google's PageSpeed Insights API, which executes a Lighthouse audit on the URL you provide. It does not estimate speed from markup — it asks Google to actually analyze the page and then reports what comes back. You can run the audit with a mobile or desktop strategy, with mobile used by default because that is how Lighthouse simulates the more constrained device profile.
The headline is the performance score out of 100. Below it the lab metrics table lists the Lighthouse measurements: First Contentful Paint, Largest Contentful Paint, Total Blocking Time, Cumulative Layout Shift, Speed Index, and Time to Interactive.
The lab metrics come from a single simulated Lighthouse run, so they describe one controlled load rather than your real users. When Google has enough real-world data for the page or origin, the response also includes a Field data section from the Chrome User Experience Report, showing how actual visitors experienced metrics like LCP, INP, CLS, and TTFB, each rated Fast, Average, or Slow.
When Lighthouse identifies meaningful savings, a Top opportunities list appears with the largest items first, such as reducing unused JavaScript or properly sizing images.
Because every check spends a call against Google's PageSpeed Insights API, the tool can be rate-limited or quota-limited, especially during busy periods, and a request can return an error or time out before Lighthouse finishes. If that happens, waiting a moment and trying again usually resolves it. Scores also genuinely vary between runs: Lighthouse simulates load conditions, so two back-to-back tests of the same URL can differ by several points without anything changing on the page. The lab metrics describe one controlled load, not a guarantee of what every visitor sees.
You submit a URL and an optional mobile or desktop choice. That URL is passed to Google's PageSpeed Insights API, which fetches and audits the page; the results are returned to you and are not stored here. Only publicly reachable URLs can be tested, since Google must be able to load the page anonymously.
Input: https://example.com/article (strategy: mobile)
Output: Performance 74/100 · LCP 2.9 s · CLS 0.04 · TBT 210 ms. Top opportunities: reduce unused JavaScript and serve images in next-gen formats.
A mid-70s mobile score with a 2.9 s LCP is decent but has clear headroom in the listed opportunities.
Input: https://example.com/ (strategy: desktop)
Output: Performance 96/100 · LCP 1.1 s · CLS 0.00. Field data (Chrome UX Report): LCP FAST, CLS FAST, INP AVERAGE.
Lab and field agree the page loads well, though the AVERAGE INP hints responsiveness could still improve.
Input: https://example.com/ (run during heavy usage)
Output: An error message indicating PageSpeed Insights returned an error or took too long to respond, with a prompt to try again.
This is the rate-limit path; retrying after a short pause normally succeeds.
Directly from Google's PageSpeed Insights API, which runs a Lighthouse audit on your URL. The tool reports Google's results; it does not compute its own performance estimate.
Lighthouse simulates network and CPU conditions for each run, so results naturally fluctuate. Run the test a few times and consider the range rather than fixating on one score.
Each run consumes a PageSpeed Insights API call, which can be rate-limited during heavy use, and Lighthouse occasionally takes longer than the request window. Waiting briefly and retrying usually clears it.
Lab metrics come from the single simulated Lighthouse run shown in the report. Field data, when available, comes from the Chrome User Experience Report and reflects how real visitors experienced the page over time.
Test whichever matters most to your audience, and ideally both. The strategies use different simulated devices, so a page often scores noticeably higher on desktop than on mobile.
The Chrome User Experience Report only has data for pages and origins with enough real Chrome traffic. New or low-traffic URLs often have none, so the field-data section simply does not appear — that is expected, not a failure.
It surfaces Lighthouse's largest estimated savings, biggest first, such as reducing unused JavaScript or properly sizing images. The items are suggestions ranked by potential impact, not a guarantee of exact time saved.
Our editorial and data-source notes explain how each check works and where estimates come from.