Keep your reports
Why it matters
Found, trusted, contacted.
Every new customer takes the same three steps before they call you. The audit checks what can stop them at each one.
01 Found
Google can crawl and index the page, and it loads fast enough on a phone that people wait for it.
Indexing · speed · mobile · local02 Trusted
A visitor, or an AI answer, can tell who runs the site: a secure connection, a real business name, reviews and contact details.
Trust · security · AI search03 Contacted
One tap to call, a button that says what it does, and a form short enough to finish.
Conversion · accessibility01 / Evidence
Other tools give you an opinion.
We give you the reading.
Every check publishes its threshold. Open any issue and you get the measured value, where we found it, and the exact change. The rows below are from an example report.
What we read
largest-contentful-paint element: img.hero__bg (1920×1080, 812 KB) render-delay: 2.9 s format: JPEG, unoptimised
measured on / · Moto G4 profile · 4× CPU throttle
The fix
Serve the hero as AVIF at 2× the rendered box (960px), preload it, and drop the render-blocking font stylesheet above it. Modelled LCP after the change: 2.2 s.
effort medium·impact high
What we read
<link rel="canonical" href="http://example.com/pricing"> → 301 https://example.com/pricing → 301 https://example.com/pricing/
11 pages affected · found in <head>
The fix
Write canonicals as absolute HTTPS URLs with the final trailing-slash form. Consolidates link equity that is currently bleeding across three hops.
effort low·impact high
What we read
GET /blog/technical-seo-basics <meta name="description"> … absent SERP fallback: first 158 chars of body copy
9 of 25 crawled URLs
The fix
Generate 140–158 character descriptions from each page's first paragraph, then hand-edit the three templates that drive real traffic.
effort low·impact medium
What we read
document outline: /checks h2 → "Crawlability" h2 → "Metadata" h1 → (none)
/checks, /privacy, /report/*, /login
The fix
Promote the leading H2 to an H1 on each template. Keeps the outline valid and gives the crawler an unambiguous topic per page.
effort low·impact medium
What we read
schema.org validation Organization … valid SoftwareApplication … valid BreadcrumbList … valid (25/25)
25 URLs parsed
The fix
Nothing to do. Adding aggregateRating once you collect reviews would make you eligible for star markup in the SERP.
effort —·impact —
02 / Coverage
One score. 279 checks. 11 groups.
The groups every report is built from. The counts here are an example report's — the SEO checker page names every test by group, and the full check list includes the ones no crawler can measure.
Content & AI answer format
1 failed2 warnings
example report
Trust & E-E-A-T
0 failed1 warning
example report
Conversion & contact paths
0 failed1 warning
example report
On-page SEO
3 failed2 warnings
example report
Site setup: robots, sitemap & tracking
1 failed1 warning
example report
Speed optimisations
1 failed3 warnings
example report
Server & security
0 failed2 warnings
example report
Mobile usability
0 failed1 warning
example report
Advanced SEO
2 failed3 warnings
example report
Accessibility
2 failed2 warnings
example report
Local & multi-region SEO
1 failed1 warning
example report
02.1 / What is measured
The checks behind the score, with the thresholds we hold you to.
Each threshold is the one freecheck enforces, linked to the body
that published it rather than asserted here.
Core Web Vitals, measured not estimated
Three numbers decide whether a page feels fast, and Google publishes the threshold for each: LCP 2.5 s or better (how long until the biggest thing on screen appears), CLS 0.1 or lower (how much the layout jumps while loading), and INP 200 ms or better (how quickly the page answers a tap). We read them from PageSpeed Insights for both mobile and desktop, so the value in your report is the one Google recorded — not a score our own machine invented. The thresholds are web.dev’s, quoted as published.
The things that stop a page being indexed at all
Speed only matters once a page can be found. A crawl follows your internal links the way
a search engine does and reports what blocks them: a robots.txt rule that
disallows the page, a noindex left in a template after launch, a canonical tag
pointing somewhere else, redirect chains, links that 404, and pages that exist but appear
in no sitemap. Each is reported with the URL it was found on and the response we got, so
there is nothing to reproduce before you can act on it.
Structured data, parsed and checked against the type
JSON-LD is read from every crawled page and validated against
Google’s
own list of supported types — required properties that are missing, @id
references that point at nothing, and effort spent on rich results Google has since retired
(the sitelinks searchbox, for one). Markup that is merely present is not the same as markup
that earns a result, and the report says which one you have.
What a number cannot decide
Some of what matters is judgement, not arithmetic: whether a title matches what the page is really about, whether headings describe the content, whether a page answers the question it ranks for. Those checks are read rather than counted, and they are the part of the report a threshold could never produce. Why we built it this way covers the reasoning, and the FAQ answers what a scan does and does not touch on your server.
03 / Pipeline
Six stages, one pass over your site.
A single page comes back in about a minute. A full crawl is roughly twenty, almost all of it the writing of the report — we email you when it lands. Performance is measured, not modelled: field Core Web Vitals plus a Lighthouse lab run.
01 Fetch
We request your pages exactly the way a search engine crawler does — same headers, same patience.
02 Render
Headless Chrome with JavaScript on, so we see the page a visitor actually gets.
03 Extract
Metadata, resources, headings, links, schema and every on-page element worth checking.
04 Measure
279 checks across content, SEO, speed, security, mobile, accessibility and more.
05 Score
Each value is held against a published threshold. Pass, warn or fail — never a vibe.
06 Prioritise
Sorted by modelled impact over effort, so the top of the list is worth doing first.
04 / The report
What the audit tool hands you.
Sections of a real site-crawl report, unedited — delivered as a web page and a PDF — next to what those problems do to a visitor. Every card names the page it was read on, the value, the threshold, and the fix.
05 / Questions
Before you run it.
The short answers. The full FAQ covers what we store, what a crawl does to your server, and pricing.
Is the SEO audit tool really free?
Yes. 5 single-page scans a month with no account and no card. The free scan runs 50 threshold checks — title, indexability, structured data, Core Web Vitals, HTTPS, mobile — and shows the measured value for each. A paid report crawls the whole site and adds the checks a threshold cannot decide.
What does the checker analyze?
The served HTML, the response headers, robots.txt, the sitemap, the page rendered in headless Chrome, structured data, real-user Core Web Vitals from the Chrome UX Report, Lighthouse lab performance, TLS, Safe Browsing, AI-crawler access and llms.txt, and accessibility with axe-core. 279 checks in all; every one is listed with its threshold.
How is this different from other SEO checkers?
Every finding shows the number we measured and the published threshold it was held against, not a score with no explanation. The check total is fixed per tier and the score is arithmetic, so two runs of the same page score the same. Each issue comes with the fix and a ready-to-paste prompt for a developer or a coding agent.
How long does an audit take?
A single page comes back in about a minute. A site crawl of up to 25 pages takes roughly twenty minutes, most of it writing the report — we email you when it is ready.
Does the scan affect my website?
It fetches your pages the way a search-engine crawler does — a few dozen requests for a single page, one page at a time on a crawl. It changes nothing, logs in nowhere and submits no forms. The scanner IP is published so you can allow it through a firewall.
06 / Compare
€19 once, not €199 a month.
What the price difference actually buys, and what it doesn't. An agency does the fixes for you on a retainer; the audit gives you, or them, the exact list. Pricing.
Find out what's holding your rankings back.
5 free single-page scans. No account, no card, no sales call.