audit_speed

Cost: 15 credits.

audit_speed measures how fast your pages actually load, using Google Lighthouse run by DataForSEO. Unlike the other audits it does not read a crawl — you name the page URLs and each one gets its own measurement — so no crawl_site run is needed first. It is synchronous: the table comes back immediately, with no background job to poll.

What it does

Pass between one and five page URLs. Each is measured on its own, and each page block reports:

• Performance score — Lighthouse's own 0–100 score for the page.
• Lab Core Web Vitals and friends — First Contentful Paint, Largest Contentful Paint, Speed Index, Total Blocking Time, Cumulative Layout Shift, and Time to Interactive, each printed with Lighthouse's own formatting.
• The biggest opportunities — the improvements Lighthouse estimates the largest load-time saving for, largest first, with the estimated milliseconds saved.

These are lab measurements. Lighthouse loads the page once, on the vendor's machine, under simulated throttling. That is a repeatable diagnostic, not a record of what your visitors experienced — the field metrics Google reports from real Chrome users (including Interaction to Next Paint) are a different measurement, and this tool does not claim them.

A metric Lighthouse did not produce for a page gets no line at all, and a page it could not score says so in words. Nothing is filled in with a zero: on a speed report a fabricated zero would read as the best possible news.

URLs are canonicalized before anything is charged. A bare domain is read as its https home page, the same page named twice is measured once, and an address that is not a public http(s) page — localhost, an internal hostname, a file: or javascript: URL — is refused for free.

Who can run it

audit_speed needs a paid credit balance. Despite sitting in the audit family it reads live data from a paid third-party provider — one real browser run per page — so it is not available on trial credits. Buy any credit pack and it unlocks straight away; your existing credits keep working for crawls, the other audits, reports and Search Console tools.

If live DataForSEO access is unavailable on this deployment, the tool returns a clear "page speed measurements are not yet enabled on this deployment" message and charges you nothing — no credits are reserved or spent. SeoGrep never returns sample or placeholder figures dressed up as a real measurement.

Example

Ask your MCP client in plain language:

How fast is example.com's home page and pricing page?

Then hand a finding straight back to your assistant:

What would it take to fix the biggest opportunity on the pricing page?

Input

Returns

A heading naming how many pages were measured and that the figures are lab measurements; then one block per page — the URL, when it was measured and by which Lighthouse version (and where it redirected to, if it did), the performance score, the metric lines, and the opportunity list.

An empty list, more than five URLs, and any address that is not a public web page are all rejected before anything is charged; while live data is off you get the "not yet enabled" message instead — also free.

Billing

One call is one flat price, charged once, whether you measure one page or five. Behind it, each page is its own Lighthouse run; if any of them fails, the whole call fails and you are not charged — a partial table is never billed.

Limitations

Results are not stored. Each call returns its measurement to the conversation and nothing else keeps it, so there is no speed history in the dashboard yet and no "compared with last week" — run it again to get a fresh measurement.

The five-URL cap per call is deliberate: a Lighthouse run is a real browser loading a real page, which is why this is measured per page rather than across a whole site.