Bulk PageSpeed Insights Checker: Core Web Vitals & Lighthouse

Check the speed of hundreds of pages in one run with Google's official PageSpeed Insights API (v5). You get one clean row per URL and device:

• Lighthouse scores (0-100): performance, accessibility, best practices, SEO
• Lab metrics: LCP, CLS, TBT, FCP, Speed Index, Time to Interactive and server response time (TTFB)
• Real-user data (Chrome UX Report / CrUX) when Google has it: LCP, INP, CLS, FCP and TTFB at the 75th percentile, with Google's FAST / AVERAGE / SLOW rating, for the URL and for the whole origin
• Core Web Vitals verdict: PASSED or FAILED, using Google's thresholds (LCP ≤ 2.5 s, INP ≤ 200 ms, CLS ≤ 0.1)
• Top opportunities: the biggest performance problems on the page, with estimated savings in ms / KB and the metrics they affect
• Monitoring mode: compare with the previous run and flag regressions. Works well with an Apify schedule.

Mobile, desktop or both. Sitemaps are supported.

What does Bulk PageSpeed Insights Checker do?

Bulk PageSpeed Insights Checker runs Google PageSpeed Insights (Lighthouse) on a list of URLs and returns performance, accessibility, best practices and SEO scores, lab metrics (LCP, CLS, TBT, FCP, Speed Index) and real-user Core Web Vitals from the Chrome UX Report (CrUX), for mobile and desktop. Use it for site speed audits, Core Web Vitals monitoring and performance regression alerts.

Who is it for

• SEO agencies and freelancers auditing client sites or prospect lists
• Web developers checking a whole site after a release
• E-commerce teams watching product and category pages every week
• Lead generation: find businesses with slow websites and offer them help

How to use it

1. Paste your URLs (one per line) or add a sitemap URL.
2. Pick the device (mobile, desktop or both) and the Lighthouse categories.
3. Optional but recommended for more than ~200 checks: paste your own free Google API key (see below).
4. Run. Export the results as JSON, CSV, Excel or via API, or connect them to Google Sheets, Make, Zapier, etc.

Input example

{
  "urls": ["https://www.wikipedia.org/", "https://web.dev/"],
  "strategy": "both",
  "categories": ["performance", "accessibility", "best-practices", "seo"],
  "maxOpportunities": 5,
  "compareWithPreviousRun": false
}

Output example (one row, shortened)

{
  "url": "https://web.dev/",
  "strategy": "mobile",
  "status": "ok",
  "performanceScore": 49,
  "accessibilityScore": 90,
  "bestPracticesScore": 100,
  "seoScore": 92,
  "lab": { "lcpMs": 4056, "fcpMs": 3906, "cls": 0.117, "tbtMs": 1142, "speedIndexMs": 4947, "ttiMs": 12174, "ttfbMs": 3 },
  "coreWebVitals": "FAILED",
  "coreWebVitalsSource": "url",
  "fieldData": { "overallCategory": "SLOW", "lcpMs": 3057, "inpMs": 226, "cls": 0.01, "fcpMs": 3062, "ttfbMs": 1740, "...": "categories per metric" },
  "originFieldData": { "...": "same fields for the whole site" },
  "opportunities": [
    { "id": "render-blocking-insight", "title": "Render-blocking requests", "displayValue": "Est savings of 2,550 ms", "score": 0, "estimatedSavingsMs": 2550, "estimatedSavingsKb": 0, "affectedMetrics": ["LCP", "FCP"] }
  ],
  "previous": null,
  "regression": null,
  "lighthouseVersion": "13.5.0",
  "analysisTime": "2026-10-26T16:04:30.282Z"
}

Every row has exactly the same keys. fieldData is null when Google has no real-user data for that exact URL (common for small sites); originFieldData may still be there. Pages that cannot be analysed (site down, blocked, timeout) come back as rows with status: "failed" and the reason in error, and they are not charged.

The dataset has three views: Scores & Core Web Vitals, Real-user data (CrUX) and Top opportunities. A RUN_SUMMARY record in the key-value store shows how many checks succeeded, failed, were retried or skipped.

Your own Google API key (free, 2 minutes)

The Actor works out of the box with a shared key, limited to 200 checks per run and shared by all users (Google allows ~25,000 checks per day per key). For bigger lists, use your own free key:

1. Open https://console.cloud.google.com/ and create a project (no billing account or credit card needed).
2. Go to APIs & Services → Library, search PageSpeed Insights API and click Enable.
3. Go to APIs & Services → Credentials → Create credentials → API key. Restrict it to the PageSpeed Insights API.
4. Paste the key in the Your Google API key input. It is stored as a secret input and never written to the dataset or the log.

With your own key you can check up to 10,000 URL × device combinations per run and run up to 8 checks in parallel.

Monitoring and regressions

Turn on Compare with previous run and schedule the Actor (daily or weekly). Each row then includes previous (score and metric changes since the last run) and regression: true when the performance score dropped by at least the threshold you set (default 10 points). History is kept in your own named key-value store pagespeed-history.

Note: lab scores naturally move a few points between runs. Use a threshold of 10 or more to avoid noise.

Pricing

Pay per event: USD 4 per 1,000 analysed URLs (USD 0.004 per URL and device). "Both" devices counts as two analyses. Failed URLs are free. Apify's platform usage is included.

Limits and good behaviour

• Uses only Google's official API. Your target sites are loaded by Google's Lighthouse servers, not by this Actor.
• Each check takes 10-40 seconds at Google. Parallel checks: up to 3 with the shared key, up to 8 with your own key.
• Automatic retries with back-off on rate limits (HTTP 429) and Google server errors (5xx). If the daily quota of the key is used up, or the key is invalid, the run stops cleanly and tells you why.
• The run respects your max charge per run: it never analyses more URLs than you will pay for.
• If no URL could be analysed, the run is marked as failed (never a silent empty success).

Why use Bulk PageSpeed Insights Checker

• One row per URL with both lab and field data: Lighthouse scores, LCP, CLS, TBT, FCP, Speed Index, TTFB, plus real-user Core Web Vitals (CrUX) for the URL and the origin, and the PASSED / FAILED verdict.
• Bulk input: a list of URLs and/or sitemaps, mobile, desktop or both.
• Regression monitor: with compareWithPreviousRun each row shows the score change and a regression flag when it drops by more than your threshold.
• Works without a Google key for up to 200 analyses per run; add your own free key for up to 10,000.
• Failed URLs are not charged.

Cost examples

Prices at the current pay-per-event rates (see Pricing). You can set a maximum charge per run in the run options.

Job	Approximate cost
100 URLs, mobile	about $0.40
100 URLs, mobile + desktop	about $0.80
Weekly check of 50 URLs (mobile), one month	about $0.80

Use it from your code, an integration or an AI agent

Python (pip install apify-client):

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("latamdatatools/pagespeed-bulk-checker").call(run_input={
    "urls": [
        "https://www.wikipedia.org/",
        "https://web.dev/"
    ],
    "strategy": "mobile"
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

JavaScript / Node.js (npm install apify-client):

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('latamdatatools/pagespeed-bulk-checker').call({
  "urls": [
    "https://www.wikipedia.org/",
    "https://web.dev/"
  ],
  "strategy": "mobile"
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

No code:

• Schedules: run it daily or weekly from Apify Console (Schedules), with the same input saved as a Task.
• Integrations: send results to Google Sheets, Slack, email, Make, Zapier, n8n or Airbyte from the Integrations tab, or call your own webhook when a run finishes.
• AI agents: add it as a tool in Claude, Cursor, VS Code and other MCP clients through the Apify MCP server (https://mcp.apify.com), so the agent can run Bulk PageSpeed Insights Checker and read the results.

FAQ

Why are my scores different from pagespeed.web.dev? It is the same API and the same Lighthouse, but lab tests vary between runs (network, server load). Field data (CrUX) is the stable, real-user number.

Why is fieldData empty? Google only publishes real-user data for pages and origins with enough Chrome traffic.

Can I check pages behind a login? No. PageSpeed Insights only loads public pages.

Do I need a Google API key? No for up to 200 analyses per run. For bigger lists, create a free PageSpeed Insights key (see the section above) and paste it in apiKey. The key is stored as a secret input.

Can I run it on every deploy? Yes. Call it from your CI with the Apify API (see the code above) and fail the pipeline if a row comes back with regression: true.

Questions or a missing field? Open an issue in the Issues tab.