One run
Lighthouse, lab
When PageSpeed is available, Lighthouse loads the page under simulated conditions and provides diagnostic suggestions.
Performance, HTTP
Check a public page’s performance
Enter a URL, choose mobile or desktop, and run a test. See the available performance measurements and HTTP checks, with no account required.
No accountNo cardPublic URL
Data sources
Start with the source of the result.
A report may combine a browser test, real-user data and HTTP checks. Each source answers a different question.
Previous 28 days
CrUX, real users
Data from Chrome users appears when there are enough samples. It may cover the URL or the whole origin.
Public URL
HTTP checks
The test checks the server response, headers, static HTML and resource samples. It does not run the page’s JavaScript.
A missing metric means it was not measured. It does not mean the page passed.
How PageSpeed describes its data sourcesMetrics
What each metric tells you.
LCP, INP and CLS are Core Web Vitals. FCP and TTFB help investigate what caused a problem. Metric availability depends on the data source.
LCP
Largest Contentful Paint
Time until the largest element in the viewport appears. It is usually a hero image or a large heading.
- Good
- ≤ 2.5 s
- Warn
- 2.5–4 s
- Poor
- > 4 s
INP
Interaction to Next Paint
Time from a click or other interaction to the next interface update. INP replaced FID as a Core Web Vital in March 2024.
- Good
- ≤ 200 ms
- Warn
- 200–500 ms
- Poor
- > 500 ms
CLS
Cumulative Layout Shift
A measure of unexpected layout shifts while someone uses the page. A high CLS means content moves on screen.
- Good
- ≤ 0.1
- Warn
- 0.1–0.25
- Poor
- > 0.25
TTFB
Time to First Byte
Time from sending an HTTP request to receiving the first response byte. It includes DNS, TLS and server processing time.
- Good
- ≤ 800 ms
- Warn
- 800–1800 ms
- Poor
- > 1800 ms
FCP
First Contentful Paint
The point when the browser shows the first content element, such as text, an image or an SVG.
- Good
- ≤ 1.8 s
- Warn
- 1.8–3 s
- Poor
- > 3 s
These thresholds follow the PageSpeed Insights documentation. TTFB is a diagnostic guide; one HTTP run does not describe every visitor’s experience.
From result to cause
Where to look when a page is slow.
Choose one signal from the report and confirm its cause before editing code. After deployment, repeat the test for the same URL and device.
LCP
The largest element appears late
Check the image size, when it starts downloading, and whether more requests are ahead of it.
INP
Interactions feel slow
Inspect long JavaScript tasks and the specific interaction. A load-only test cannot measure response after a click.
CLS
Elements shift on screen
Check image dimensions, reserved ad space and content inserted after the first render.
TTFB
The server responds late
Check redirects, caching and HTML generation time. This stage comes before content can render.
When to use it
When to run another test.
Save the earlier result yourself. The free Auditor has no history. For a useful comparison, test the same URL and device view.
After a release
Check selected pages and compare publicly detectable signals with the previous result.
Before a campaign
Review technical risks on key pages before increasing traffic.
After a refactor
Compare results before and after the change to identify possible regressions.
Public-signal comparison
Check publicly available indicators for selected pages using the same scope.
On this public tool the PageSpeed key belongs to CometWeb, so you do not configure your own. When the key is available, the audit combines Lighthouse with CometWeb HTTP checks; without it, HTTP checks remain. The report always identifies its data source.
Audit duration depends on the selected mode, the analyzed page response, external measurement-service availability and the number of resources. The interface shows a waiting state during the test. The completed report identifies its data source.
No. This form does not accept a login, password or user session. If the URL leads to a login screen, the test will not inspect the content behind it. Do not put access tokens in the URL.
HTTP mode does not run a browser and evaluates different signals from Lighthouse. With PageSpeed, measurement conditions, viewport, throttling and available CrUX field data can still affect results. Compare runs that use the same scope and configuration.
The URL is sent to the CometWeb server for the audit. The free Auditor does not expose project history or a project profile. Current processing rules and product scope are described in the legal documents and information shown with the form.
Colors group a result against thresholds used by that measurement: green marks the highest range, while yellow or red identifies more findings to review. The score describes detected technical signals; it does not measure traffic, sales or conversions.
Run an audit after a significant release, before an important campaign, or when the page or its dependencies change. Choose a monitoring frequency based on release cadence, page importance and regression risk within the analyzed scope.
Comparison
CometWeb Auditor and alternatives
Each tool answers a different question. This is a comparison of scope, not a ranking or a run-time benchmark.
History and alerts depend on the product and plan. The free Auditor has neither, so see CometWeb Insight for working with results, priorities and follow-up verification.
This tool patches one problem at a time
Core Web Vitals in every audit, with priorities, history and before/after comparison.