Understand what is slowing your website down.

Google considers Largest Contentful Paint good under 2.5 seconds, Interaction to Next Paint good under 200 milliseconds, and Cumulative Layout Shift good under 0.1. A page passes Core Web Vitals when 75 percent of real visits meet all three thresholds. Pulse measures the same metrics in a controlled browser, which tells you whether the page is capable of hitting those numbers before real traffic decides for you.

PageSpeed Insights reports Lighthouse scores and, where Google has enough traffic data, field measurements from the Chrome UX Report. Pulse runs the same Lighthouse engine but adds its own analysis on top: what the page is actually made of, which images are shipped larger than they are displayed, how much JavaScript is never executed, what third party code costs you, and a ranked list of fixes with the measured saving attached to each one.

Lighthouse scores performance from five timing metrics and nothing else. Pulse weighs those same timings but also accounts for total page weight, request count and third party bytes, because a page can hit good timings on a fast connection while shipping several megabytes that punish a real visitor on a phone. Both numbers appear side by side on every report, and the breakdown shows exactly where each point was lost.

Lab data. Every Pulse audit is one scripted page load, from one region, on an emulated device, in a real headless Chromium instance. That makes results repeatable and comparable between runs, which is what you want when checking whether a fix worked. It is not a record of what your visitors experienced, and Pulse never presents it as one.

Interaction to Next Paint measures how quickly a page responds after somebody clicks, taps or types. A scripted page load performs no interactions, so there is no genuine INP to report. Rather than invent a number, Pulse marks the metric unavailable and shows Total Blocking Time instead, which is the closest signal a lab run can give you about main thread responsiveness.

Most pages finish in 25 to 60 seconds. The audit launches a browser, runs Lighthouse, inspects the rendered DOM, executes accessibility checks with axe-core and fetches robots.txt, so it is slower than a page load but far more thorough. Heavy pages with large JavaScript bundles or WebGL content take longer, particularly on the mobile profile, which applies four times CPU throttling.

Pulse runs axe-core inside the rendered page against WCAG 2.1 level A and AA, and reports every violation with the elements that trip it and the rule that was broken. Automated testing reliably catches around a third of accessibility problems. A clean result is a good sign rather than a certificate, and keyboard navigation and a screen reader pass still need a person.

No. Pulse audits publicly reachable pages over http or https. Addresses on private networks, localhost and cloud metadata endpoints are refused, and a page that returns 401 or 403 is reported as blocked rather than audited. This is a deliberate security boundary, because the service accepts arbitrary URLs from anyone.

Start at the top of the report. Findings are ranked by measured impact rather than by category, so the first item is the one returning the most bytes or milliseconds on your page. In practice oversized images are the most common largest win, followed by JavaScript that is downloaded and parsed but never executed during load.

Yes. Audits are free and no account is required. Usage is rate limited per IP address so that one visitor cannot occupy the browsers everyone else is waiting for, and your audit history is kept in your own browser rather than on a server.