PageSpeed Insights

PageSpeed Insights runs Google’s Lighthouse audit on a page and adds field data from real Chrome users. It answers: how fast and how well built is this page, and what is holding it back? Slow or broken pages are harder for search crawlers and AI engines to read.

Add your API key

The integration runs on your own Google API key, so audits use your Google quota. The PageSpeed page has the same steps under How to get your API key:

Go to console.cloud.google.com and create a project, or select an existing one.
Open APIs & Services → Library, search for PageSpeed Insights API and click Enable.
Open APIs & Services → Credentials and click Create Credentials → API key. Copy the key.
In Rankry, paste it into API key under PageSpeed Insights API key and click Save key.

Each project uses its own key, so rate limits are separate for each project. Google’s default quota for the PageSpeed Insights API is 25,000 requests a day per Google Cloud project. If you hit limits, restrict the key to the PageSpeed Insights API under API key restrictions in Google Cloud.

Rankry stores the key encrypted. To remove it, click Disconnect at the top of the page.

Run an audit

Enter a page URL, choose mobile or desktop, and click Run audit. Desktop is selected when you open the page. The report opens as soon as Google returns it.

Each page keeps its latest report per device. Running the audit again replaces it.
Running the same audit again within a minute returns the report you already have.
In read-only mode the button reads Read-only and audits can’t run.

Reading the report

The top line says when the report ran. The header has two halves:

Determine your site’s real-world performance: field data from real Chrome users, with Google’s overall rating and how each metric is spread across visits. A page without enough Chrome traffic shows No field data yet. When only your whole site has enough traffic, the report shows origin-wide data and says so.
Identify performance issues: lab scores from this run, as gauges for Performance, Accessibility, Best Practices and SEO, plus Agentic Browsing when Google returns it. Scores read Poor (0–49), Needs improvement (50–89) or Good (90–100).

Below the header:

Metrics: the lab timings, such as First Contentful Paint, Largest Contentful Paint, Total Blocking Time, Cumulative Layout Shift and Speed Index, each with a rating, plus how the run was set up.
Screenshots of the page as it loaded.
Performance audits, grouped as Google groups them. Use Show audits relevant to to narrow them to one metric, or click a metric chip in the header. Audits that passed sit under Passed audits.
A section for each other category, with its failing audits, Additional items to manually check, Passed audits and Not applicable.

Using it with other tools

AI tools connected to the MCP server can read your saved reports and run new audits with this project’s key. Rankry Brain can read the reports but doesn’t run audits.