PageSpeed & Lighthouse Bulk Audit + AI Readiness

Audit hundreds of pages at once with Lighthouse 13 – the engine behind Chrome DevTools and PageSpeed Insights – and get a clean, spreadsheet-ready row for every page and device:

• Scores for Performance, Accessibility, Best Practices and SEO (0–100), plus AI readiness – Lighthouse 13's new Agentic Browsing score: how well AI agents can read and use the page
• Lab metrics: FCP, LCP, TBT, CLS, Speed Index, server response time
• Core Web Vitals verdict (good / needs improvement / poor)
• Top fixes with estimated time and size savings (e.g. Reduce unused JavaScript – 1.7 s, 1.3 MB)
• Diagnostics: page weight, number of requests, DOM size, main-thread and JavaScript time
• Failed accessibility, SEO and best-practice checks in plain language
• Real-user Core Web Vitals (CrUX): LCP, INP, CLS at the 75th percentile
• Optional full HTML report per audit – the same report you know from Chrome DevTools, ready to share with developers
• A short problem list per page for quick filtering

Two engines – pick what suits you

	PageSpeed Insights engine (recommended)	Local Lighthouse engine
What you need	Your own free Google API key	Nothing
Where the audit runs	Google's PageSpeed Insights servers	A real Chrome inside the Actor
Numbers match pagespeed.web.dev	Yes – same servers and settings	Close, but see Good to know
Real-user data (CrUX)	Included	Included when you add a key

Auto mode (default) uses PageSpeed Insights when you enter a key and the local engine otherwise.

How to get a free Google API key (about 5 minutes)

1. Open Google Cloud Console and create a project (no billing or credit card needed).
2. APIs & Services → Library: enable PageSpeed Insights API and Chrome UX Report API.
3. APIs & Services → Credentials → Create credentials → API key. Under API restrictions select just those two APIs.
4. Paste the key into the Google API key field. It is stored encrypted and used only for your own runs.

Google's free quota is 25,000 PageSpeed Insights audits per day per project.

AI readiness (Agentic Browsing)

Lighthouse 13 adds a category that checks whether AI agents and AI browsers can work with a page: a well-formed accessibility tree, a valid llms.txt, WebMCP tools and forms, and a stable layout. You get the score and the failed checks for every page – useful as more traffic and purchases come through AI assistants. Remove Agentic browsing from the categories if you don't need it.

Typical uses

• SEO and web agencies: audit all client sites, or every template of a large site, in one run.
• Before/after releases: compare scores and Core Web Vitals across deployments.
• Monitoring: schedule a weekly run and track trends in Google Sheets, Looker Studio or any dashboard.
• AI readiness: see which pages AI agents struggle with (accessibility tree, llms.txt, WebMCP).

Input example

{
  "urls": ["https://apify.com", "example.com/pricing"],
  "strategies": ["mobile", "desktop"],
  "categories": ["performance", "accessibility", "best-practices", "seo", "agentic-browsing"],
  "googleApiKey": "AIza...your key...",
  "saveReports": true
}

Output example (shortened, real result)

{
  "url": "https://apify.com/",
  "strategy": "mobile",
  "engine": "psi",
  "scores": { "performance": 38, "accessibility": 100, "bestPractices": 73, "seo": 85, "agenticBrowsing": 98 },
  "metrics": { "fcpMs": 2701, "lcpMs": 14026, "tbtMs": 1718, "cls": 0.002, "speedIndexMs": 6714 },
  "labCoreWebVitals": { "lcp": "poor", "cls": "good", "tbt": "poor", "pass": false },
  "opportunities": [
    { "id": "unused-javascript", "title": "Reduce unused JavaScript", "savingsMs": 1800, "savingsKb": 1309, "affects": ["LCP"] },
    { "id": "bootup-time", "title": "Reduce JavaScript execution time", "savingsMs": 1650, "savingsKb": null, "affects": ["TBT"] }
  ],
  "diagnostics": { "mainThreadMs": 5803, "jsExecutionMs": 3384, "totalByteWeightKb": 5388, "requests": 243, "domElements": 1639 },
  "failedAudits": {
    "accessibility": [],
    "bestPractices": ["Uses third-party cookies", "Browser errors were logged to the console"],
    "seo": ["Links do not have descriptive text", "robots.txt is not valid"]
  },
  "field": {
    "level": "url", "collectionPeriod": "2026-09-28..2026-10-25",
    "lcp": { "p75": 2427, "rating": "good" }, "inp": { "p75": 395, "rating": "needs-improvement" },
    "cls": { "p75": 0.02, "rating": "good" }, "coreWebVitalsPass": false
  },
  "problems": ["low_performance_score", "low_bestPractices_score", "low_seo_score", "poor_lab_lcp", "poor_lab_tbt", "fails_core_web_vitals_field"],
  "reportUrl": "https://api.apify.com/v2/key-value-stores/…/records/report-0001-mobile"
}

Good to know

• Lab scores vary between runs on any tool, PageSpeed Insights included (±5–10 points on script-heavy pages is normal). Compare trends with the same settings.
• Local engine: Lighthouse 13 with the standard mobile (mid-range phone, slow 4G, simulated throttling) and desktop presets. Accessibility, SEO and best-practice results match PageSpeed Insights. On script-heavy pages, Total Blocking Time usually reads higher than in PageSpeed Insights because cloud CPUs are shared – use the PageSpeed Insights engine when you need numbers that match Google's tool.
• Each local audit starts a fresh browser profile (no cache or cookies carried over), just like PageSpeed Insights.
• One audit takes roughly 10–40 seconds (PageSpeed Insights sometimes up to 1–3 minutes; 6 audits run in parallel).
• Memory is set automatically: 1 GB for the PageSpeed Insights engine, 8 GB for the local engine (Lighthouse needs CPU, and on Apify more memory also means more CPU). The local engine needs at least 4 GB.
• Real-user data exists only for pages and sites with enough Chrome traffic; otherwise the Actor falls back to the whole origin or says no data is available.
• Some websites block automated visits (HTTP 403). Such audits are reported as failed and are not charged.

Pricing

Free during the launch period – you only pay Apify platform usage. A pay-per-audit price is planned for 2027; any change is announced in Apify Console in advance. Failed audits and invalid inputs are never charged.

FAQ

Why is my score different from PageSpeed Insights? With the PageSpeed Insights engine you get Google's own result. With the local engine, hardware and network location differ, which mostly affects Total Blocking Time on heavy pages.

Is my API key safe? The field is a secret input: Apify stores it encrypted and the Actor uses it only to call Google's PageSpeed Insights and CrUX APIs for your run. It never appears in the output.

Is this legal? The Actor loads the pages you provide, exactly like a visitor would, or asks Google's official API to do so with your key. It does not log in or bypass any protection.

Support

Built and maintained by VEmKa (Czech Republic, EU). Questions or ideas? Open an issue on the Issues tab – support in English and Czech.

---

PageSpeed Insights, Lighthouse and Chrome are trademarks of Google LLC. This Actor is independent and not affiliated with or endorsed by Google.