On-page SEO is everything about a page that you control from the page itself: whether Google may index it, how it appears in results, what the content says and how it is structured, where it links, how it behaves on a phone, and what it tells machines about itself. The checker above reads the page and its robots.txt, runs about thirty checks, and links each problem to the tool that goes deeper.

The six groups of checks

Indexing and crawl

Whether Google is allowed to see and keep the page: HTTPS, a noindex in the robots meta tag or the X-Robots-Tag header, a Disallow in robots.txt that matches the URL, the canonical tag, redirects, and the page language. Any failure here outranks everything below it — a perfect page that says noindex does not rank.

Search appearance

What a searcher sees before clicking. The title, measured in pixels the way Google cuts it; the meta description; the URL; the favicon Google shows on mobile. With a target keyword, whether it appears where searchers look.

Content

One H1 that names the subject, headings in order without skipped levels, enough words to answer the question, the keyword early and used naturally, and alt text on images. Word count is counted from the main content, without navigation and footers, so a template-heavy page is not credited for its menu.

Links

Internal links are how Google finds pages and learns which ones relate. A page that links to almost nothing is a dead end; internal links marked nofollow starve your own pages. The internal link checker goes link by link.

Mobile and speed

A responsive viewport, which Google’s mobile-first indexing depends on; how long the HTML took to arrive; its size; whether it is compressed; and scripts in the head that block rendering. These are the signals visible in one request. They are not a substitute for real-user speed data.

Social and structured data

Open Graph and X card tags, which decide how a shared link looks, and JSON-LD structured data, which makes a page eligible for rich results. The schema validator checks the markup against what Google requires.

Fix in this order

1. Anything that stops indexing. noindex, a robots.txt block, a page not served over HTTPS.
2. The title and the H1. The two strongest statements of what the page is. Check the title’s width with the SEO title checker.
3. The content itself. Does it answer the query completely, with a structure a reader can scan?
4. Internal links. Into the page from related pages, and out to them.
5. The description, social tags and structured data. They change how the result looks and how often it is clicked.
6. Speed signals. Worth fixing, rarely the reason a page does not rank.

What a page checker cannot tell you

• Whether the page can rank. That depends on the competition for the query and on the site’s authority — links, which no on-page check sees.
• Real-user speed. Core Web Vitals come from Chrome users; one request from a server is a signal, not a measurement of experience.
• Problems across pages. Duplicate titles, orphan pages and cannibalisation only show when many pages are compared.
• Content rendered by JavaScript. The checker reads the HTML as served. If a page shows its text only after scripts run, the word count here will be low — and that is worth knowing, because many crawlers read it the same way.

About keyword density

There is no ideal percentage, and there never was. The checker flags a keyword that never appears in the text and one that makes up more than about three per cent of it, where it starts to read unnaturally. Between those, write for the reader.

Frequently asked

Do I need a score of 100?

No. The score is a quick read of how many checks pass. Some warnings are deliberate — a short page can be the right page — and one critical issue matters more than ten warnings.

Why does my page show almost no words?

Its content is probably rendered by JavaScript after the HTML loads. Google renders pages, eventually; many other crawlers, link previews and AI fetchers do not.

Does the checker store my page?

No. It fetches the page when you ask, keeps the result for ten minutes so a repeated check is instant, and nothing else.

How is this different from Search Console?

Search Console reports what Google has already seen, days later, for pages it knows about. This checks any page right now, including one that is not live in Google yet.