Free website grader

What this website grader checks

Enter the address of any public page. Our server loads it the way a desktop browser would. It first follows the redirects, for the address you typed and for its http:// version. Then four requests run side by side on the final page: one reads its HTML, one reads the site's robots.txt and tests the page against it, one tests up to 100 links on the page, and one fetches the icons the page declares. The results are combined into 22 checks, each with a weight, and the weights add up to 100.

The checks are the ones that decide whether a page can be found and shared at all. A single noindex tag or one line in robots.txt can keep a page out of Google no matter how good it is, so those weigh most. Titles, descriptions and the mobile viewport come next, then broken links, share images and icons. Each check in the report says what we found, why you lost points and what to change, with a link to the tool that shows the details.

It grades one page at a time, usually the home page or a page that should bring in leads. Grade your most important pages one by one: a site is only as healthy as the pages people land on.

How the score is calculated

Each check ends in one of four states. A pass earns the full weight, a partial result earns half, a fail earns nothing, and a check that could not run (for example because the site refused our link test) is left out, so the score is worked out from the checks that did run. The score is the points earned divided by the points available, times 100, rounded to a whole number.

One rule overrides the sum. A page that does not load, carries noindex, or is blocked for Googlebot in robots.txt is capped at 49, whatever else it gets right, because none of the other checks matter until Google can read and keep the page. Grades follow the usual school scale: A from 90, B from 80, C from 70, D from 60, F below 60.

The weights are our judgment of what costs a typical business site the most, based on Google's own documentation (see the sources). They are not Google's ranking factors: no tool outside Google knows those, and any grader that claims to is guessing.

The problems that cost the most

Most sites lose points in the same few places. In rough order of damage:

• A forgotten noindex. Sites are often built on a staging address with "discourage search engines" switched on, and launched without switching it off. The page looks perfect to visitors and never appears in Google. Google's documentation is explicit: a page with noindex is dropped from results the next time it is crawled.
• A robots.txt that blocks too much. A Disallow: / left over from development, or a rule meant for one folder that matches the whole site, stops Googlebot from reading the page. A robots.txt that answers with a server error is worse: Google temporarily treats the whole site as disallowed while the errors last.
• Redirect chains and http:// leftovers. Every hop slows the first visit and makes the move harder for search engines to follow. The http:// version of every address should go to https:// in one permanent redirect, and internal links should point straight at the final address.
• Missing or duplicated titles. The title is the blue link in Google. Without one, or with the same title on every page, Google writes its own, and it is rarely the one you would choose.
• Broken links. A link to a page that no longer exists is a dead end for visitors and for crawlers. Google says 404s on their own do not hurt the rest of the site, but they waste the visit and whatever authority the link was meant to pass on.
• Blank share previews. Without Open Graph tags and a share image, a link posted on LinkedIn, Facebook or WhatsApp shows a grey box or a random picture from the page. For many businesses that is the first impression.

For the full detail of each area, open the dedicated tool linked from every check: the meta tag checker shows every tag with link previews, the broken link checker lists every link with its status, the redirect checker traces each hop, and the robots.txt checker tests any URL and crawler against the rules.

Speed and Core Web Vitals: why there is no speed score here

Many graders put a speed score at the top. A fair one needs either a real browser loading the page on a throttled phone connection, which is what Lighthouse in PageSpeed Insights does, or field data from real Chrome users, which is what Google uses for Core Web Vitals. A quick server check can do neither, so this grader does not score speed at all rather than show a number it cannot back up.

What the report does show, without scoring it, is how long your server took to answer our request from Cloudflare's network. A slow answer there (over about a second) usually points to slow hosting or a page that is built on every visit without caching. For the real measurement, every report has a button that opens Google PageSpeed Insights for the same address.

Website grader or a full SEO audit?

A grader is a quick health check of one page against fixed rules. It is the right tool to catch the expensive mistakes after a launch, a redesign or a plugin update, and to check that a fix worked. It cannot tell you whether your content answers what your buyers search for, how your site compares with competitors, which pages Google actually keeps in its index, or why traffic dropped last month. That needs Google Search Console data, a crawl of the whole site and someone who reads the results with your business in mind.

If you are comparing graders: HubSpot's Website Grader, the best known, describes itself as scoring performance, mobile readiness, SEO and security. This one leaves speed to PageSpeed Insights and goes deeper on what can quietly keep a page out of Google or make it share badly, and it shows the weight of every check so you can see exactly why you lost points.

Sources