What the free site check reads on your site

How a result is built

• It reads your home page and some of the pages it links to.
• Each finding is critical, warning or info. The score starts at 100 and each finding takes some off, the most serious the most.
• The free check shows the three most serious findings and counts the rest.
• No language model reads your site. Each finding comes from a fixed rule applied to the HTML it fetched.

The checks

28 checks in 6 areas, filed as the result card files them. 5 of them look across pages, so they only fire once the check has read a second page.

Metadata

Give your homepage a title tag

The page has no <title> in its head.Search results and browser tabs show a bare URL instead of your product name.title_missing

Cut your title to what search results show

A title longer than 60 characters.Search results cut long titles, so the end of yours may not show.title_too_long

Write a meta description for your homepage

No meta description.Google writes its own snippet for your listing instead of your pitch.meta_description_missing

Resize your meta description to 120 to 160 characters

A meta description under 120 or over 160 characters.A short one wastes the space and a long one gets cut.meta_description_length

Add a canonical link to your homepage

No canonical link.Duplicate or parametered versions of a page can split your search ranking.canonical_missing

Fix the canonical link that points nowhere

A canonical that cannot resolve: an unfilled placeholder such as /project/[slug], a relative address or an unreadable URL.It sends search engines to a URL that does not exist.canonical_broken

Keep one canonical link per page

More than one canonical link on a page.When they conflict, search engines may ignore all of them.canonical_duplicate

Declare UTF-8 so your characters render

Garbled characters, or literal \uXXXX escapes, in the title, the meta description or the Open Graph title or description.Accented characters show as garbage in search results, tabs and shares.text_encoding_broken

Serve your homepage as HTML

A page served with a Content-Type other than text/html.Browsers and crawlers trust that header over the bytes, so the markup may show as plain source or be skipped.content_type_not_html

Give each page its own title

Two or more of the pages it read share one title.Search engines cannot tell them apart and usually show only one.duplicate_title, needs a second page

Give each page its own meta description

Two or more of the pages it read share one meta description.The snippet under each of them says the same thing.duplicate_meta_description, needs a second page

Page content

Give the page one H1 that states the outcome

No <h1> on the page.Neither readers nor crawlers get one clear headline for what the page is about.h1_missing

Keep one H1 per page

More than one <h1> on a page.A single h1 should state the one thing the page is about.h1_multiple

Give the page enough text to rank

Fewer than 150 words of visible text on the home page.It reads as thin to search engines and gives AI crawlers little to cite.thin_content

Add an H1 to the pages that have none

At least three pages read, and at least 30 percent of them have no <h1>.The h1 tells a search engine and a screen reader what a page is about.h1_missing_multi, needs a second page

Accessibility

Declare the language of the page

The <html> tag has no lang attribute.Screen readers and translators cannot tell what language the page is in.html_lang_missing

Describe your images in alt text

Images with no alt attribute. An empty alt, for a decorative image, counts as set.Screen-reader users and image search lose them.image_alt_missing

Link previews

Complete your link preview tags

Any of og:title, og:description, og:image or og:url missing.Links to your site unfurl as a bare URL on social networks, Slack and iMessage.open_graph_incomplete

Add a Twitter card so shared links unfurl

No twitter:card tag.X shows a plain link instead of a large image card.twitter_card_missing

AI readiness

Add structured data so search engines can quote you

No JSON-LD structured data on the page.Search engines and AI answer engines cannot reliably identify what your product is or cite it.structured_data_missing

Remove the noindex tag blocking search engines

A robots meta tag containing noindex or none.It tells search engines to leave the page out of results.robots_noindex

Remove the noindex header from your hosting config

An X-Robots-Tag response header containing noindex or none.The same effect as the meta tag, set in the server's headers instead of in the page.x_robots_noindex

Serve your main content in the HTML

Under 60 visible words in the server HTML, and a framework mount node: Next.js #__next, #root, #app, data-reactroot, data-server-rendered or an Angular ng-version.Googlebot renders JavaScript, but AI answer engines and link-preview bots often do not, so they may see an almost empty page.client_rendered_shell

Keep utility pages out of search results

A login, signup, search, admin or dashboard page that answers 200 without noindex, found because one of your own pages links to it.Search engines can index your login and admin screens beside your real pages.utility_page_indexable, needs a second page

Shorten the redirects between your pages

Internal URLs that pass through two or more redirects before they answer.Each hop costs every visitor and crawler a round trip, and some crawlers stop before the end.redirect_chain, needs a second page

Site setup

Add a viewport tag so mobile renders correctly

No viewport meta tag.The site renders zoomed out and is hard to use on a phone.viewport_missing

Add a favicon so your tab is recognisable

No favicon link.Browser tabs, bookmarks and history show a blank icon beside your name.favicon_missing

Serve your site over HTTPS

A page served over http instead of https.Browsers label the site "Not secure".not_https

If your site is a single-page app

A React, Vue or Angular site, or a Next.js site that renders only in the browser, can send an almost empty page and fill it in with JavaScript. The check flags this as a warning, not a failure: Googlebot renders JavaScript, so the site can still be indexed. The risk is the readers that do not. AI answer engines and link-preview bots often read the HTML only. The remedy is to server-render or pre-render the first screen of copy.

What a finding looks like

Three findings, word for word as the check wrote them on 2026-10-21.

What the free check does not read

• robots.txt, sitemap.xml and a made-up address that should return 404. 8 checks need them, so 6 of the 14 checks in AI readiness run on the free check. A clean AI readiness row means nothing failed among those 6, not that nothing is wrong.
• Anything JavaScript adds after the page loads.
• Page speed and Core Web Vitals, backlinks and rankings. None of them is one of its checks.
• Login, admin and similar pages, unless one of your own pages links to them. It does not go looking for them.

What an account adds

A free account keeps the check of your site and opens the rest: every finding, not only the top three; the fix for each, with a snippet in HTML or Next.js form filled in from your own page; the robots.txt, sitemap and missing-page checks the free version skips; and a re-check after you ship a fix.

Tools that go further

• Link preview checker

  See the card X, LinkedIn and WhatsApp build from your link, and what to fix.

• AI crawler checker

  Check whether ChatGPT, Claude and Perplexity may read a page under your robots.txt.