Free tools / AEO Content Grader
Free tool · SEO content checker for AI answers
AEO Content Grader
Paste a draft or a page URL and we grade it on the five points we hold our own posts to: lists where the content is a list, headings phrased as the buyer's question, links to outside sources, a real FAQ, and prose without the machine tells. Then the structure rules our content engine checks before anything publishes.
Sample result
What a result looks like
A short draft we wrote for this page, graded live. Every line is the tool's own output.
Meets 3 of the 5 points. Start with links to outside sources.
How it works
Five points, measured
We read the page's shape
Paragraphs, list items, H2 and H3 headings, links and the FAQ, from your HTML or your markdown.
We grade five points
Each gets pass, warn or fail against a stated target, with the measurement beside it.
We run the structure rules
One H1, no skipped heading levels, each section answering its heading early, real anchor text, alt text on images.
We name the first fix
The verdict points at the first point you miss, so the rewrite starts in the right place.
Reference
The five points and our targets
| Point | We measure | Our target |
|---|---|---|
| Lists | Share of words in list items; paragraphs over 80 words | 10% or more in lists, no paragraph over 80 words |
| Question headings | Share of H2 and H3 headings ending in a question mark | Half or more |
| Outside sources | Distinct other domains linked | 3 or more |
| FAQ | Entries and median answer length | 3 or more entries, answers of 40 words or more |
| Prose | Prose score out of 50, stock phrases, em dashes | 35 or more, no stock phrases, no em dashes |
What to do with a result
What each result means, and what to do about it
| If the result says | It means | Do this |
|---|---|---|
| Lists: warn or fail | Enumerations are buried in long paragraphs. | Turn real lists into lists, and split paragraphs past 80 words. Never invent a list. |
| Question headings: fail | No heading matches the way a buyer asks. | Rephrase H2s as the questions they answer. Keep a statement when it carries its own answer. |
| Outside sources: warn or fail | Claims stand on your word alone. | Link the publisher of each figure in the same sentence. The Stat and Source Checker finds the gaps. |
| FAQ: warn or fail | No FAQ, or answers too thin to quote. | Add at least three questions buyers ask that the page does not already answer, each answered in 40 words or more. |
| Prose: warn or fail | Stock phrases or em dashes give the copy away. | Run the AI Slop Checker to see each one in place. |
Where do these five points come from?
From our own blog. In September 2026 our founder set five rules for every post and we measured the whole corpus against them. Two were broken nearly everywhere: most posts put under 10% of their words in lists, and most had no question headings at all. The grader is the measurement we ran on ourselves, pointed at your page.
The targets are our editorial standard. No engine publishes a threshold it rewards, and the grader doesn't pretend one exists. What the five points share is that each makes a passage easier to lift whole: a question heading matches the query, a list holds the parallel facts, a linked source lets a reader check the number.
This checks your site. It cannot tell you whether AI names you.
A free scan puts real buying queries to all five AI engines and shows who they name instead of you.
Check your AI visibility free →FAQ
Frequently asked questions
What is AEO content?
Content written so an answer engine can lift a passage and quote it: a heading that matches the question, an answer in the first lines under it, facts in plain text with their sources, and an FAQ that covers the follow-ups. AEO stands for answer engine optimization.
Are these thresholds what Google or ChatGPT reward?
No engine publishes a threshold. These are the targets we hold our own posts to, and we label them that way. Use them as a standard to write to, and measure the effect with a scan before and after.
Why question headings?
A buyer types a question into an assistant. A heading phrased as that question makes the passage under it easy to match. Keep a declarative heading when it states its own answer, like 'The right cadence: monthly'.
Why does it want lists?
Parallel facts in a paragraph are hard to quote one at a time. A list holds them apart. Only turn real enumerations into lists; a list invented for its own sake reads worse.
How does it find my FAQ?
From FAQPage markup first, then from disclosure lists on the page. In pasted text, a question heading followed by its answer counts as an entry.
Can I paste markdown?
Yes. Use # for headings, - for list items and [text](url) for links. Plain prose works too; it has no headings or lists to count.
How is this different from the GEO Content Score?
The GEO Content Score reads the prose of one live page against a measured sample of 290 B2B pages. The grader checks structure: lists, headings, sources and FAQ, plus the prose floor. Run both on a page you care about.
Is this free?
Yes, no signup. Our paid plans write drafts that pass these checks against the queries your buyers ask five AI engines, then measure whether those engines name you.