Performance score
The Lighthouse performance score is a weighted summary of synthetic lab metrics. Treat it as a diagnostic benchmark, not a ranking guarantee.
Website Speed Test
Test a public page with a mobile or desktop Lighthouse run. Review performance scores, lab metrics, real-user field data when available, optimization opportunities and exact diagnostics.
Mobile + desktopLighthouse lab dataCore Web Vitals when availableNo signup
Queued
Mobile performance testPreparing the performance queue.
0s
- 1AcceptedRequest received
- 2QueuedWaiting for browser
- 3MeasuringLoading and tracing
- 4ProcessingBuilding report
- 5ReadyResults displayed
Heavy Lighthouse work runs in one isolated worker. This page receives small status updates instead of holding one long server request.
Performance report
Website performance
Real-user experience
Core Web Vitals field data
Shown only when sufficient field data is available.
Controlled test
Lab performance metrics
These measurements come from the selected Lighthouse device strategy and can vary between runs.
Optimization
Performance opportunities
Potential improvements are ordered by estimated time or transfer savings when Lighthouse provides them.
Diagnostics
What is slowing or complicating the page
Technical findings that help explain the measured result.
Page weight
Requests and transfer size
Resource summary from the Lighthouse run.
| Resource | Requests | Transfer |
|---|
Passed audits
How to interpret this test
Understand the result
Mobile and desktop speed are different tests
A page can perform well on desktop and still struggle under mobile CPU and network conditions.
Core Web Vitals
Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift describe loading, responsiveness and visual stability using real-user data when available.
Lab metrics
FCP, LCP, Total Blocking Time, CLS and Speed Index help reproduce performance problems under a controlled test profile.
Opportunities
Use estimated savings to prioritize work such as reducing render-blocking resources, image payloads, unused code or excessive main-thread work.
What counts as good Core Web Vitals?
LCP ≤ 2.5 seconds
Largest Contentful Paint measures when the main visible content finishes rendering. The good threshold is 2.5 seconds or less at the 75th percentile.
INP ≤ 200 milliseconds
Interaction to Next Paint measures responsiveness to user interactions. The good threshold is 200 milliseconds or less at the 75th percentile.
CLS ≤ 0.1
Cumulative Layout Shift measures unexpected visual movement. The good threshold is 0.1 or less at the 75th percentile.
How to improve a slow website
Start with the largest measured bottlenecks rather than chasing a perfect score. Optimize the largest above-the-fold images, remove or defer unnecessary JavaScript, reduce render-blocking CSS, improve cache policy and compression, and investigate slow server response. Re-test after each meaningful change so you can confirm the improvement under the same device strategy.
Field data vs lab data
Field data reflects aggregated real-user experiences when enough data exists for the tested page or origin. Lab data is a fresh synthetic Lighthouse run. A newly launched or low-traffic page may have useful lab results but no field dataset yet.
Related website and SEO tools
Data source and limitations
This tool runs an isolated Lighthouse lab test for the selected mobile or desktop strategy. When a supported CrUX/PageSpeed provider source is configured, real-user field data may also be shown. Performance varies with page content, server conditions and test environment, so compare repeated runs and focus on persistent bottlenecks rather than a single number.