About

QA engineering tools,
not another dashboard.

That line is a constraint, not a slogan. It rules things out, which is the only useful kind of positioning.

Most site-quality tools sell you a place to look at your problems. You get a login, a graph that trends sideways, a weekly digest you stop reading in a month, and a number that goes up when someone else fixes something. The work of finding out what is actually wrong gets replaced by the work of maintaining an account.

The scanner is the opposite bet. You paste a URL, wait about twenty seconds, and get a list of specific, named defects with a severity on each one. No signup. Nothing to configure. If the list is empty, you close the tab and get on with your day — and that is a good outcome, not a failed conversion.

What that rules out

ruled out

Scores you cannot audit

Every deduction maps to a named check with a fixed weight, and the whole ledger is published. A number nobody can reconstruct is a vibe, not a measurement.
ruled out

A signup wall in front of the value

The gate sits on the three worst findings, after you have already seen the score and everything else. If the free scan is not useful on its own, the gate is doing the wrong job.
ruled out

Advice the tool cannot back up

A check either asserts something verifiable — this certificate expired, this link returns 404, this layout shifts — or it does not ship. There is no “consider improving your content strategy” tier.
ruled out

Findings you have to interpret

A finding names the thing and where it is. Titles are the defect, not the category the defect belongs to.

How it actually works

no magic

A scan reads the page you enter and asks Google’s PageSpeed Insights to render it, which is where the Core Web Vitals and the phone-viewport verdicts come from — their Lighthouse run, on their hardware, reproducible at pagespeed.web.dev. Alongside that it requests outbound links directly to see which ones are dead, reads the response headers, inspects the TLS certificate, and checks the domain’s SPF and DMARC records. Unlock a crawl and a real headless browser walks up to ten more pages. Fifty-eight assertions in total.

Then it subtracts. Every site starts at 100 and each finding costs points by severity. There is no curve. A site that scores 62 has, quite literally, thirty-eight points of named damage you can read line by line.

Where this goes

now

One tool, done properly

The website scanner. It is free, it is anonymous, and it is the thing the rest of this gets built on.
next

Reports that outlive the process

Today a report is deleted from the server the moment it reaches your browser, and lives only in that tab. Persistence is the next thing worth building, because a score is far more useful as a line than as a point.
later

More tools, same shape

Whatever comes after should also be something you can point at a URL and get an answer from in under a minute, without an account. If it needs onboarding, it is the wrong tool.

Run a scan on your own site.

Takes about 20 seconds. No account, no card.

Run a scan