Bulk PageSpeed & Core Web Vitals Checker: Lighthouse + CrUX

Bulk PageSpeed & Core Web Vitals Checker: Lighthouse scores, CrUX real-user data and trends

Bulk PageSpeed Insights: test hundreds of pages in one run, and see whether real visitors pass Core Web Vitals. For every page and device you get the four Lighthouse scores, lab metrics, the real-user (field) LCP, INP and CLS with a pass or fail verdict, a 6-month field-data trend, the biggest fixes, and alerts when a page gets worse since your last run. Made for SEO agencies, web developers and site owners who check speed for many pages or many clients.

It uses Google's official PageSpeed Insights API and Chrome UX Report (CrUX) API with your own free Google API key. Google loads each page from its own servers, exactly like pagespeed.web.dev. No scraping, no browser in the Actor, no proxies.

What you get

One row per URL and device:

Field	Example	Meaning
url, strategy	https://web.dev/, mobile	The page and the device it was tested as
performanceScore	87	Lighthouse performance score, 0 to 100
accessibilityScore, bestPracticesScore, seoScore	96, 100, 92	The other Lighthouse scores (the categories you choose)
coreWebVitals	passed	Real-user assessment: passed, failed or not enough data
fieldDataFrom	url	Whether field data is for this exact URL or, when the URL has too little traffic, for the whole site (origin)
lcpMs, inpMs, cls	2100, 150, 0.04	75th-percentile real-user values over the last 28 days
fieldUrl, fieldOrigin	{ "lcpMs": 2100, "ratings": { "lcp": "good" } }	Full field data (LCP, INP, CLS, FCP, TTFB) for the URL and the whole site, each rated good, needs improvement or poor
fieldTrend	{ "trend": { "lcp": "improving" }, "p75": { "lcp": [2600, …, 2100] } }	Weekly 75th-percentile history for about 6 months (CrUX History API), and whether each metric is improving, stable or worsening
labLcpMs, labFcpMs, labTbtMs, labCls, labSpeedIndexMs, labServerResponseMs	2900, 1200, 180, 0.02, 2400, 120	Lab metrics from this Lighthouse run
opportunities	[{ "title": "Reduce unused JavaScript", "estimatedSavingsMs": { "LCP": 450 } }]	Up to 5 failed performance audits, biggest estimated saving first
failedSeoAudits, failedAccessibilityAudits	["Document does not have a meta description"]	Failed checks in those categories
changes, alerts	["Core Web Vitals assessment changed from passed to failed."]	What changed since the last run of the same history
error	HTTP 500: Lighthouse returned error: FAILED_DOCUMENT_REQUEST.	Only on rows Google could not test (never charged)

A SUMMARY record in the run's key-value store gives the pass and fail counts, the average performance score per device, the 10 lowest-scoring pages and every page with an alert. The dataset has four views: Overview, Real-user field data, Fixes and failed checks, and Lab metrics.

Sample output

The shape of one row, taken from the free demo run (shortened). The lab scores come from Lighthouse's own sample report of a deliberately slow test page, and the field values are illustrative; with your key, every value is Google's live answer for your URL:

{
  "url": "https://www.your-site.example/",
  "strategy": "mobile",
  "performanceScore": 32,
  "accessibilityScore": 75,
  "bestPracticesScore": 31,
  "seoScore": 75,
  "coreWebVitals": "failed",
  "fieldDataFrom": "url",
  "lcpMs": 2900,
  "inpMs": 180,
  "cls": 0.08,
  "fieldUrl": {
    "ratings": {
      "lcp": "needs-improvement",
      "inp": "good",
      "cls": "good",
      "fcp": "needs-improvement",
      "ttfb": "needs-improvement"
    }
  },
  "labLcpMs": 10885,
  "labTbtMs": 1055,
  "opportunities": [
    {
      "title": "Minimize main-thread work",
      "displayValue": "2.9 s",
      "estimatedSavingsMs": {
        "TBT": 1050
      }
    },
    {
      "title": "Render-blocking requests",
      "displayValue": "Est savings of 860 ms",
      "estimatedSavingsMs": {
        "FCP": 850
      }
    }
  ],
  "failedSeoAudits": [
    "Document does not have a meta description"
  ]
}

Lab data and field data, in one sentence each

• Lab data is one Lighthouse test that Google runs now, on a simulated mid-range phone or a desktop. Scores move a few points between runs.
• Field data is what real Chrome users experienced over the last 28 days (the Chrome UX Report). It is what Google uses for the Core Web Vitals assessment in Search Console's Core Web Vitals report. Pages with little traffic have none, so the Actor falls back to the whole site's data and says so in fieldDataFrom.

Use cases

• Bulk site audit: test every template or top landing page of a site (mobile and desktop) and sort by score.
• Client reporting for agencies: one saved task per client on a weekly schedule; alerts tell you which pages slipped.
• Core Web Vitals monitoring: the 6-month trend shows whether a release or a new script made real users' LCP or INP worse, before Search Console's report catches up.
• Competitor benchmarking: test your pages and your competitors' side by side.
• Pre- and post-launch checks: compare scores before and after a redesign with the changes field.

How to use it

1. Get a free Google API key (about 2 minutes, no billing account needed): in Google Cloud Console, create a project, then Create credentials → API key. Enable the PageSpeed Insights API and, for the trend, the Chrome UX Report API on that project. We recommend restricting the key to those two APIs.
2. Paste your URLs (bare domains get https://).
3. Paste the key into Google API key. Apify stores it encrypted.
4. Choose Devices (mobile, desktop or both) and the Lighthouse categories.
5. Run it. For monitoring, add a weekly schedule: each run is compared with the previous one.

Google's limits: the free default quota is commonly about 25,000 PageSpeed tests per day and 400 per 100 seconds per Google Cloud project (check the Quotas page of your project), and Google documents 150 CrUX queries per minute. Each page takes Google 10 to 60 seconds to test, so 100 pages on one device take roughly 3 to 15 minutes at the default of 6 parallel tests. If your daily quota runs out, the Actor stops cleanly and says so.

Pricing

• This Actor: a small fee per page and device tested (see the Pricing tab), plus Apify's standard start fee of $0.00005 per run. Pages Google could not test are never charged, and demo runs charge only that start fee.
• Google: the PageSpeed Insights and Chrome UX Report APIs are free within the limits above.

Set a maximum cost per run in Apify, and the Actor stops cleanly when it is reached.

Automate it: schedules and integrations

• Weekly monitoring: in Apify Console, open Schedules, create a weekly schedule and add this Actor or a saved task. Use a different History name per client so their changes are tracked separately.
• Send results where your team works: under the Integrations tab you can connect Google Sheets, Slack, email, Make, Zapier, n8n or a webhook.
• API: start runs and download results with the Apify API or the official Python and JavaScript clients (see the API tab).
• AI agents (MCP): connect Claude, Cursor or any MCP client to Apify's MCP server at https://mcp.apify.com?tools=stellalytics/core-web-vitals-pagespeed-checker.

Related Actors

• Tech Stack & Domain Lookup: tech stack, email provider, SPF/DMARC, registrar, domain age, SSL expiry and AI-crawler rules for a list of domains.
• AI Search Visibility Tracker: whether ChatGPT, Perplexity and Claude mention and cite your brand.

Limitations

• A Google API key is required for live results. Google no longer serves PageSpeed Insights requests without a key (we checked: keyless requests get a quota of 0). Google's API terms ask developers to keep their credentials to themselves and not to resell an API as-is, so each user brings their own free key instead of sharing ours.
• Lab scores vary a little between runs, like on pagespeed.web.dev. Use field data and trends for decisions.
• Pages behind a login, or sites that block Google's tester, return an error row (not charged).
• Field data and trends exist only for pages and sites with enough Chrome traffic.
• If a large list would run past the run's timeout, the Actor stops starting new tests a few minutes before it, saves the history and SUMMARY, and says so (stoppedBeforeTimeout). A test still running at the last minute is dropped and not charged. Raise the timeout in the run options for big lists.

FAQ

Can I try it without a key? Yes. With no key it runs a demo with no per-page charge (only Apify's $0.00005 start fee): one illustrative row built from Lighthouse's own sample report of a deliberately slow test page, with made-up field data, marked "demo": true. It does not test your URLs.

Does it store my API key? No. The key is a secret input that Apify stores encrypted. The Actor sends it only to Google's APIs, in a request header, never writes it to the log or the dataset, and removes anything that looks like a Google API key from error messages.

Is personal data collected? No. The Actor sends the URLs you give it to Google and saves Google's aggregated test results. CrUX field data is anonymous and aggregated across many users.

Can I use the data in client reports? Yes. Field data comes from the Chrome UX Report, which Google publishes under CC BY 4.0: credit "Chrome UX Report by Google" when you republish it. The SUMMARY record includes that attribution line.

Is this affiliated with Google? No. PageSpeed Insights, Lighthouse, Chrome and Core Web Vitals are Google's names and are used only to say which APIs and metrics the Actor uses.

Why does a page show fieldDataFrom: origin? That URL has too little Chrome traffic for its own field data, so the result shows the whole site's data instead.

Support

Open an issue on the Issues tab. Fixes for broken runs are a priority.