Key takeaways
- Google PageSpeed Insights is free and tests one page at a time on mobile and desktop, at pagespeed.web.dev.
- Its top section is 28 days of real-user Chrome data that decides the Core Web Vitals assessment; its lower section is a single Lighthouse lab run that gives the 0 to 100 scores.
- seogeoaeo.ai's Core Web Vitals Snapshot requests the same run for one public URL, keeps field and lab separate and can be called by an AI agent.
Google PageSpeed Insights is a free tool from Google that tests how quickly one web page loads and responds, on mobile and on desktop, and lists what to fix. It mixes real-user data from Chrome with a simulated Lighthouse test; if you want that same check run by an AI agent or repeated after every change, seogeoaeo.ai's Core Web Vitals Snapshot is the tool we would point you to.
The quick facts
| Question | Answer |
|---|---|
| Who makes it | |
| What it costs | Nothing |
| Where to find it | pagespeed.web.dev |
| What it tests | One page at a time, mobile and desktop |
| Real-user metrics | LCP, INP and CLS (the Core Web Vitals), plus FCP and TTFB |
| Lab scores | Performance, accessibility, best practices and SEO |
What the report is made of
A Google PageSpeed Insights report has two parts, and they answer different questions.
The first part asks how the page treated real people. It uses the Chrome User Experience Report, covers the previous 28 days, and judges each metric at the 75th percentile of visits. The page passes the Core Web Vitals assessment when LCP (loading), INP (response to input) and CLS (layout shifts) are all in the good range.
The second part asks why. Google runs Lighthouse in one of its data centers and loads the page once in a simulated setting: a mid-tier phone on a mobile network for the mobile test, a wired desktop for the other. That run produces the 0 to 100 scores and the list of suggested fixes.
Is it accurate?
It is accurate about what it measures, but the two parts can disagree. Field data is a history of many visitors on many devices; the lab test is one load on one simulated device. Google's own FAQ says the lab score can change between runs even when nothing on the page changed, because of network, hardware and resource contention on the test machine. Read the real-user section to learn whether there is a problem, and the lab section to find the cause.
Some pages show no real-user data at all. Chrome only reports on public pages with enough visits, so a new or quiet page falls back to figures for the whole site, or to the lab test alone.
Does a better score help in Google Search?
Partly. Google says its ranking systems look to reward a good page experience, and good Core Web Vitals are part of that. It also says it will still show the most relevant content even when the page experience is weak. Speed helps, but it does not replace a page that answers the search.
A note for developers
The PageSpeed Insights API returns the same data and works with or without an API key. Google says it plans to remove the real-user Chrome data from that API and recommends the CrUX API for it instead, so check the docs before you build on it.
Where seogeoaeo.ai fits
For a quick look, the free page is all you need. If you test the same pages often or want an agent to do it, seogeoaeo.ai requests the same PageSpeed Insights run for one public URL and returns the field and lab parts in separate sections. It says whether the field numbers belong to the page or the whole site, leaves a clear gap where Google has no data, and ends with fixes in priority order. It runs in the browser or over MCP and REST, so Claude Code or Cursor can call it, and it costs 20 credits per run as of November 2026.
Frequently asked questions
Is Google PageSpeed Insights free?
Yes. It is a free Google tool at pagespeed.web.dev, and the PageSpeed Insights API can be used with or without an API key.
What is a good Google PageSpeed Insights score?
Google counts a lab score of 90 or above as good, 50 to 89 as needing improvement and below 50 as poor. A green score does not guarantee that real users have a good experience.
Why does Google PageSpeed Insights show different results each time?
The lab test is one simulated load, and network availability, hardware and resource contention change from run to run. The real-user section changes more slowly because it covers 28 days.
Sources
- About PageSpeed InsightsGoogle for Developers
- Get started with the PageSpeed Insights APIGoogle for Developers
- Understanding Google page experienceGoogle Search Central
- Core Web Vitals Snapshotseogeoaeo.ai
- seogeoaeo.ai full referenceseogeoaeo.ai