Lighthouse & Core Web Vitals Audit: Bulk PageSpeed Scores

Lighthouse Audit runs Google's open-source Lighthouse in a real Chrome browser against any list of URLs and returns performance, accessibility, best-practices and SEO scores plus Core Web Vitals (LCP, CLS, TBT) for every page. No Google API key, no PageSpeed Insights quota, no login: paste URLs, press Start, and get one clean row per page, with the biggest fixes and the score change since your last run.

What does Lighthouse Audit do?

• Audits up to thousands of pages in one run, mobile, desktop or both.
• Returns the four Lighthouse category scores (0 to 100): performance, accessibility, best practices, SEO.
• Returns lab Core Web Vitals and speed metrics: LCP, FCP, CLS, TBT, Speed Index, TTFB, page weight, request count and DOM size.
• Flags coreWebVitalsPass (LCP ≤ 2.5 s, CLS ≤ 0.1, TBT ≤ 200 ms).
• Lists the top speed opportunities with estimated savings in milliseconds and bytes, plus the failed accessibility, SEO and best-practice checks.
• Checks an optional performance budget (for example "performance at least 80, LCP under 2,500 ms") and outputs budgetPass and the exact violations.
• Remembers each URL's last score and outputs previousPerformanceScore and performanceScoreChange, so a scheduled run becomes a site speed monitor.

Why use it instead of the PageSpeed Insights API?

Most PageSpeed tools on the Store forward your URLs to Google's PageSpeed Insights API. That API is rate limited, needs an API key at volume, and only reaches public pages Google can see. This Actor runs Lighthouse itself (the same engine behind PageSpeed Insights and Chrome DevTools, Apache-2.0 licensed), so:

• there is no API key to create and no daily quota to hit;
• you choose the categories, so performance-only runs are faster;
• every run on the same machine type gives comparable numbers, which is what matters for tracking trends and catching regressions;
• you get budgets, change tracking and top fixes in the same row, ready for a spreadsheet, a dashboard or an alert.

Who is it for?

• SEO agencies auditing client sites or a prospect list in bulk.
• Developers and QA checking a release for performance regressions on a schedule.
• E-commerce teams watching product and category page speed.
• AI agents that need a quick, structured page-quality check for a URL.

How to run a bulk Lighthouse audit

1. Paste your page URLs into Page URLs (one per line; example.com works too).
2. Pick Device: mobile (Google's mobile-first default), desktop, or both.
3. Optionally narrow Lighthouse categories or add a Performance budget.
4. Click Start. Download results as JSON, CSV, Excel, or read them through the API.

To monitor a site, save the input as a Task and add a daily or weekly schedule. Each run fills performanceScoreChange so you can filter for pages that got slower.

Input example

{
  "urls": ["https://example.com", "https://www.wikipedia.org"],
  "device": "mobile",
  "categories": ["performance", "seo"],
  "performanceBudget": { "minPerformanceScore": 80, "maxLcpMs": 2500 },
  "compareWithPrevious": true
}

Output example

{
  "url": "https://example.com/",
  "finalUrl": "https://example.com/",
  "device": "mobile",
  "performanceScore": 100,
  "accessibilityScore": 100,
  "bestPracticesScore": 100,
  "seoScore": 80,
  "lcpMs": 913,
  "fcpMs": 913,
  "cls": 0,
  "tbtMs": 0,
  "speedIndexMs": 913,
  "ttfbMs": 16,
  "totalByteWeight": 2147,
  "requestCount": 2,
  "coreWebVitalsPass": true,
  "budgetPass": true,
  "budgetViolations": [],
  "topOpportunities": [{ "id": "render-blocking-resources", "title": "Eliminate render-blocking resources", "savingsMs": 151, "savingsBytes": 0 }],
  "failedAudits": [{ "category": "seo", "id": "meta-description", "title": "Document does not have a meta description" }],
  "previousPerformanceScore": 100,
  "performanceScoreChange": 0,
  "lighthouseVersion": "12.x",
  "error": null
}

Pages that cannot be loaded (bad URL, DNS failure, server error) return a row with error filled in and are not charged.

Use with AI agents / MCP

The Actor has one required field and works zero-shot from the Apify MCP server, the Apify API or any agent framework. Minimal input:

{ "urls": ["https://example.com"] }

The agent receives one row per page with numeric scores and a short list of fixes, small enough to reason over directly. Typical prompts: "Audit these 20 landing pages and list the ones with LCP over 2.5 seconds", or "Has our homepage performance score dropped since last week?"

How much does it cost to run a Lighthouse audit?

Pay per event: $0.02 per successful audit ($20 per 1,000), with automatic volume discounts on higher Apify plans. One URL on "both" devices counts as two audits. Failed pages are free. There is no start fee and no extra platform charge on top of the per-audit price.

Lighthouse API

You can call this Actor as a Lighthouse API: send a POST to the Apify run-sync endpoint with the input above and receive the dataset items in the response. Schedules, webhooks, Zapier, Make and n8n integrations work out of the box.

FAQ

How do I check Core Web Vitals for many pages at once?

Paste all the URLs and run. Each row has lcpMs, cls, tbtMs and coreWebVitalsPass. Sort or filter the dataset view "Scores and Core Web Vitals".

Are the scores the same as PageSpeed Insights?

They come from the same Lighthouse engine, but every machine and network is different, so lab scores vary a few points between tools and even between runs. Use one tool consistently for trends. Field data (real-user CrUX data) is not included; these are lab measurements.

Can I audit pages behind a login or on localhost?

No. The Actor audits public pages only and rejects local or private network addresses.

Is it legal?

Yes. The Actor loads the public pages you choose, the same way a visitor's browser does, and runs the open-source Lighthouse tool on them. It does not scrape Google, does not use any private API and does not collect personal data. Only audit sites you are allowed to test, and keep request volume reasonable for small sites.

How long does an audit take?

Usually 7 to 45 seconds per page, depending on how heavy the page is. Choosing fewer categories makes runs faster.

What about the TBT instead of INP?

Interaction to Next Paint needs real user interaction, so lab tools report Total Blocking Time as its closest proxy, exactly like PageSpeed Insights' lab section.

Legal and data

• Engine: Google Lighthouse (open source, Apache License 2.0) in headless Chrome.
• Data: only the public pages you list; no login, no CAPTCHA bypass, no personal-data collection.
• Change tracking stores only each URL's last performance score and timestamp in a key-value store in your own Apify account. To keep history across runs, pick a store in the optional "History key-value store" field; without it the Actor still audits every page and simply leaves the comparison fields empty if it cannot create its own store.

Alternatives and related tools

Compared with PageSpeed Insights API wrappers, this Actor needs no Google key and adds budgets and change tracking. Need job data instead? See Magnova's Career Site Jobs and Workday Jobs Actors. Questions or feature requests: open an issue on the Issues tab, and we will respond as quickly as we can.