AI On-Page SEO Analyzer

What the analyzer checks

Every audit runs the same set of checks, grouped into seven weighted categories that combine into the overall score out of 100. The weighting reflects real-world impact: technical problems and content depth move the number more than a missing canonical tag does.

• Title optimization — presence, length against the ~30–60 character range Google typically displays, and whether it is likely to be truncated.
• Meta description — presence and length against the ~120–160 character range, since a missing description hands Google the job of writing your snippet.
• Heading structure — H1 count, empty headings, duplicated headings, and whether the hierarchy skips levels (H1 straight to H3).
• Content — total word count of the visible text with scripts and styles stripped out, plus paragraph structure.
• Images — how many images carry meaningful alt text, and whether below-the-fold images use lazy loading.
• Links — internal versus external counts, nofollow links, empty anchors, and generic anchor text like “click here”.
• Technical SEO — HTTP status, noindex directives, canonical tag, viewport tag, language attribute, HTTPS, and the length of any redirect chain.

Reading your score

The gauge at the top is a weighted average of the seven category bars beneath it. A score in the 80s or 90s means the on-page fundamentals are in good shape; the 50–79 range usually means one or two categories are dragging the rest down; below 50 normally points to something structural, like a missing title or a page that returns a non-200 status.

Treat the category bars as the actionable part. A page can score 74 overall while having a perfect heading structure and a completely absent meta description — the bars tell you which of those it is, so you fix the thing that is actually broken rather than rewriting a title that was already fine.

The score is a heuristic built on well-established on-page best practices. It is not Google's ranking algorithm, and no free tool has access to that. What it does reliably tell you is whether the controllable on-page signals are present and sensibly configured.

Working through the issues list

Issues are sorted by severity — critical, high, medium, then low — and each one expands to show why it matters and how to fix it, usually with a concrete code example you can copy. Work top-down. A critical issue such as a noindex directive or a missing H1 will hold a page back regardless of how well everything else is tuned.

Critical and high issues are worth fixing before you publish. Medium issues are worth a pass when you next touch the page. Low issues are genuine improvements but rarely the reason a page is underperforming, so they belong at the bottom of the list.

Using the AI recommendations

Below the score, the AI Recommendations panel sends the measured data — your actual title text, headings, counts — to an AI model and asks for prioritized, page-specific advice. Because it works from measured values rather than guesses, the suggestions tend to be concrete: a rewritten title of the right length, a note that your six H2s cover four topics but skip the one your title promises, and so on.

The numbers themselves are never generated by AI. Word counts, link counts, and character lengths are computed from the parsed HTML; the AI layer only adds judgment and phrasing on top. If the AI service is unavailable, the measured report and rule-based issue list still work exactly as before.

What it cannot see

The analyzer fetches your page's HTML the way a crawler would, which means it sees what is in the server response. Content that only appears after JavaScript runs in the browser may not be counted — a fair reflection of what simpler crawlers see, but worth knowing if you run a single-page app.

It also audits one URL at a time rather than crawling your whole site, and it has no access to off-page data such as backlinks, domain authority, or your actual rankings. For sitewide technical issues, run the important templates through it individually — a product page, a category page, a blog post — since problems usually live in the template rather than in one page.

A practical workflow

• Run the page before you publish, and fix anything critical or high while editing is still cheap.
• Re-run it after publishing to confirm the live page matches what you intended — CMS templates often add or strip tags.
• Compare a page that ranks well against one that does not; the category bars usually show the gap quickly.
• Export the report as a text file to attach to a ticket or share with whoever owns the template.
• Re-check after any theme, plugin, or migration change, since those routinely alter canonical tags, viewport tags, and heading markup.

Fixing what the audit finds

Each category maps to a focused tool that fixes that specific problem faster than editing by hand. If the title bar scored low, run your topic through the SEO Title Generator and pick a variant inside the character limit. If the meta description is missing or the wrong length, the Meta Description Generator writes three options sized to display in full.

A low links score usually means thin internal linking, which the Internal Link Analyzer breaks down in detail — internal versus external, nofollow, and empty or generic anchor text. If the technical category flagged a redirect chain, trace it hop by hop with the Redirect Chain Checker.

For content problems, paste the page body into the Word Counter to check depth and sentence length, and into the Keyword Density Checker to confirm the page actually states its topic rather than dancing around it.

Frequently asked questions

Related tools