What the check tells you
A website answers far more than its pages. Every response carries headers that tell browsers how careful to be: whether to insist on HTTPS, which scripts to trust, whether other sites may frame the page, what to send to the next site a visitor clicks to. Most of them are invisible when everything goes well, which is exactly why they're so often missing. This check reads them the way a browser would and grades the result.
What we check
Entry points. People reach a site four ways: http:// and https://, each with and without www.
These are the site's entry points. We follow each one hop by hop and show
every redirect on the way. We check that plain HTTP ends on HTTPS on your own
site, that no chain loops or drops back from HTTPS to HTTP, that all four end on the same final address, and
that the chains are short. A www name with no address is noted, not failed.
Security headers. On the page the chains end on we read:
- HSTS, which makes browsers use HTTPS for your site for a set time. We check that it is present, valid, lasts at least a year and covers subdomains.
- Content-Security-Policy, the strongest defence against injected scripts. We flag
policies that still allow inline scripts,
evalor scripts from anywhere, and policies that only report. - Framing protection, through
frame-ancestorsorX-Frame-Options, against clickjacking. X-Content-Type-Options: nosniff,Referrer-Policy,Permissions-PolicyandCross-Origin-Opener-Policy.- Version numbers in
ServerandX-Powered-By, which tell scanners what to try.
Cookies. For every cookie the page and its redirects set, we check the
cookie flags: Secure on HTTPS, HttpOnly on session cookies, a SameSite
value, and the rules for __Host- and __Secure- names. We never keep cookie values.
Protocols. Whether the server offers HTTP/2, found with one TLS handshake, and whether it advertises HTTP/3. Both are information only.
Preload readiness. Whether the domain meets the published rules of the HSTS preload list, so browsers could use HTTPS even on a first visit.
How the grade works
The grade starts at 100 and loses points for each problem, weighted by how much it exposes visitors. No
HTTPS at all costs the most. A missing redirect to HTTPS, a redirect loop, a downgrade or missing HSTS cost
a lot. A missing CSP, a weak CSP, framing and insecure cookies cost a fair amount. Small things such as a
missing nosniff or a long redirect chain cost a little. Information findings, such as a missing
Permissions-Policy or no HTTP/2, cost nothing. A is 90 and up; below 50 is F.
Reading the result
The first block shows each entry point: where it started, every hop with its status code, and the page it ended on. A green tick means the chain ended on a working page; a cross means it failed, with the reason in words. Next comes the header table, one row per header with a plain verdict and the exact value we received, followed by preload readiness, the cookie table and the protocol line.
If none of the four entry points answers, the check returns an error and you aren't charged. When nothing listens on port 80 or 443 at all we say so plainly, and the domain report skips its website section rather than failing it.
How we fetch
We fetch as DNSSuperBot, from our own servers, and only the pages this check needs. Each hop is fetched once and checked against our safety rules again before we connect, so a redirect can't send us to a private address. We go easy on every site: at most a couple of dozen requests per check and a shared limit per host across all our users. If a site shows us a bot challenge, we report that entry point as not checked rather than trying to get around it.
What to do next
Each finding links to a fix guide with the exact header or setting for common servers and CDNs. Once you've made changes, run the check again with "fresh" ticked to skip the cache. The redirect tracer shows any single address hop by hop, and the domain report puts this grade next to your DNS, email and SSL results.