Google Lighthouse Checker

Get real Google Lighthouse scores for any URL — performance, accessibility, best practices, and SEO — plus Core Web Vitals metrics (LCP, FCP, CLS, TBT, Speed Index, TTI).

Under the hood this actor calls the Google PageSpeed Insights (PSI) API, which runs the exact same Lighthouse engine on Google's own servers and returns the results as JSON. That means you get authentic Lighthouse output without running a local Chromium browser — so runs finish in seconds and use minimal compute.

How It Works

1. You provide a list of URLs and a device strategy (mobile or desktop).
2. For each URL the actor calls Google PageSpeed Insights: Google runs Lighthouse and returns the audit.
3. The actor parses the Lighthouse result and emits one dataset row per URL with the category scores and key metrics.

No local browser, no Node.js, no Chromium — just Google's real Lighthouse data over a fast HTTP call.

Input Parameters

Parameter	Type	Description
urls	array	URLs to audit
categories	array	Any of: performance, accessibility, best-practices, seo
device	string	mobile or desktop (PSI strategy)
maxRequests	integer	Max URLs to audit (1–100)
api_key	string (optional override)	Google PSI API key. Normally leave this blank — the key is supplied by the platform through the PAGESPEED_API_KEY environment variable. Google's anonymous keyless quota is exhausted (verified: HTTP 429 "Quota exceeded for quota metric 'Queries'"), so a key is required.

Output

One row per URL, for example:

{
  "url": "https://example.com/",
  "requestedUrl": "https://example.com",
  "strategy": "mobile",
  "success": true,
  "fetchTime": "2026-06-30T12:00:00.000Z",
  "lighthouseVersion": "12.x.x",
  "source": "google-pagespeed-insights",
  "performanceScore": 100,
  "accessibilityScore": 88,
  "bestPracticesScore": 100,
  "seoScore": 90,
  "metrics": {
    "largestContentfulPaint": { "displayValue": "1.2 s", "numericValue": 1200, "score": 0.98 },
    "cumulativeLayoutShift": { "displayValue": "0", "numericValue": 0, "score": 1 }
  }
}

Scores are 0–100. Metrics carry Lighthouse's displayValue, raw numericValue, and per-audit score. Scores are never fabricated: only URLs that produced a real Lighthouse result become dataset rows. A URL that does not respond, or that Lighthouse cannot audit, is skipped, listed in the ERRORS key-value record, and not charged.

Pricing

• Result (apify-default-dataset-item): $0.10 per audited URL — one row per URL that returned a real Lighthouse result.
• Actor start (apify-actor-start): $0.00005 per event (one event per GB of run memory, minimum one).

No monthly fee. Skipped URLs, blocked runs and empty runs cost nothing beyond the start event. A 10-URL audit costs $1.00.

Notes

• Data comes straight from Google's Lighthouse via PageSpeed Insights, so scores match what you'd see at pagespeed.web.dev for the same URL.
• The API key comes from the platform. The actor reads PAGESPEED_API_KEY from the environment; the api_key input is only an override. If neither is set, the run succeeds with 0 rows and tells you which variable to set — nothing is charged.
• Fail-fast URL precheck. Each URL is probed before a PSI call is spent on it, so unreachable hosts never burn quota and never bill.
• Blocks fail loudly. If PageSpeed Insights refuses the run (quota, auth or outage) after retries, the run ends FAILED with the reason in the status message rather than finishing green with no data.
• Local Lighthouse is a fallback only, used solely when the run has at least 2048 MB of memory and the Lighthouse CLI is present. The actor runs comfortably at the 256 MB minimum on the PSI path.

Related tools

• Tech Stack Detector — BuiltWith Alternative
• Page Speed Analyzer — Lighthouse & Web Vitals
• StackOverflow Scraper — Q&A & Dev Trends
• GitHub Repo Stats — Deep Analytics