Best on page SEO checker tools compared, and the order to read their findings in
Which on page tool to use for a single page, for a hundred pages and for a brief, with the one rule that makes any of their reports useful.
Published by AI Knows Us · Updated 21 October 2026
On page tools split into three groups that get sold as one, and buying the wrong group is the most common mistake in this part of the market. Group one checks a single URL and tells you what is in it. Group two checks every URL on your site at once, which is a crawler wearing a different label. Group three compares your draft with the pages currently ranking and produces a brief, which is a different data pull again.
The criterion used to place every tool below: what does it compare your page with? A tool that compares your page with its own rulebook gives you a tidy list. A tool that compares your page with the pages above you gives you the finding that changes a position. Both are useful and only one of them is worth a monthly fee.
The three groups, and which one you actually want
| Group | Compares your page with | Buy it when |
|---|---|---|
| Single page checker | A fixed checklist: title length, one h1, alt text, internal links, structured data | You are fixing one page today and want the obvious misses in thirty seconds. Free tools cover this completely. |
| Site wide on page auditor | The same checklist, applied to every URL it can reach | You have a template problem. One fix usually resolves hundreds of findings, which is why the count looks frightening and the work does not. |
| Content optimiser | The pages currently ranking for your query | You publish weekly and want the brief before the writer starts. This is the group that earns a subscription. |
The order to read any on page report in
Almost every tool in groups one and two presents findings alphabetically or by its own severity scoring, and neither order matches what is actually blocking your page. Read them in this order instead, and stop at the first thing that is wrong:
Can Google reach and index the page? Status code, noindex, canonical, robots.txt, and whether it is in the index at all. Does the page work for a visitor on a phone? A banner covering the answer costs more than a missing alt attribute. Does the first paragraph answer the query in two sentences? Does the title match what was asked? What do the three pages above you cover that yours does not? Does anything on your own site link to it with a descriptive anchor? Then, and only then, the technical housekeeping.
The findings to read and not act on mechanically
On page tools report everything they can measure, which means some of their findings are arithmetic rather than advice. Keyword density has no target percentage to hit. Word count targets belong to the question and not to a tool default, so a definition page is finished at three hundred words. The score out of a hundred is a vendor weighting and has no relationship to position. Meta keywords are a field nobody reads. And a tool that wants every internal anchor to be the same commercial phrase is describing a pattern to avoid.
A note on scope, so nobody reads this page for the wrong thing
This page is about choosing a tool. The step by step method for analysing a single page, in eight layers with the free way to check each one, is a separate page of ours and it owns that question deliberately: see the eight checks on a single page rather than duplicating them here. Two pages answering neighbouring questions is coverage; two pages chasing the same query is not, so this one stays on the tool choice.
The on page factor that lives outside your page
Everything above is inside your HTML, where you control it. What you do not control is which businesses an assistant names when somebody asks it who to buy from, and that answer is composed per asker rather than stored, so no on page checker can read it.
AI Knows Us measures only that. The questions are the ones your buyers would type, the apps are ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews, and the output is who each answer named. Our pricing page publishes what a check costs in credits, with a credit standing for one question put to one app on one occasion and a Claude check standing at three of them, because Claude costs more to run. We sell nothing in the three groups above, so if you want an on page tool the table is your answer and we are not in it. On reach, Ahrefs publishes a wider view than ours, covering AI Mode and Copilot which we do not read at all, with its per engine cost published as 8 checks per update for Claude.
One on page element gets ignored until a link looks wrong in a chat window, and then it becomes urgent: SEO and Open Graph, the full tag set to ship and how to fix a wrong preview. We counted durable inbound links across 1,655 tracked pages on this site on 21 October 2026 and 696 of them had none at all; that page was one of them.
What people ask when choosing an on page SEO tool
Do I need a paid on page tool at all?
Not for groups one and two. View source, Search Console URL Inspection and a free crawler cover a small site completely. Pay for group three, the content optimiser, if you publish often enough that a brief saves you an hour a week.
What single on page change moves a page most often?
The opening. Make the first two sentences answer the question the page was written for, and make the title read like the thing that was asked. It takes ten minutes, needs no software, and across sites it is the change that pays off most often.
How many h1 tags should a page have?
One, describing the page in the reader's words, with h2 and h3 forming a genuine outline underneath. That outline is what a reader scans and what an assistant lifts from, so it is worth writing deliberately rather than letting a theme generate it.
Why does my page score 95 and rank badly?
Because the score measures your page alone and ranking is a comparison. Open the page above you beside yours and the reason is usually visible in five minutes: it answers more of the question, or it has links yours does not.
Should an on page tool see the same page Google does?
Not always, if your page depends on JavaScript. Search Console URL Inspection returns the rendered HTML Google holds for your URL, which is the only check that settles that argument, and it is free.