Skip to content
Surfilby Adam Hafez
Free tool

AEO checker: AI answer readiness score for articles

AEO checker that scores an article's opening answer, headings, ids, lists, citations, schema and date for AI answers. Free, runs in your browser.

Runs entirely in your browser. Nothing you enter is uploaded, logged or stored.

Key takeaways

  • Google states there are no additional requirements or special optimizations needed to appear in AI Overviews or AI Mode.
  • This score measures structure only, so a page can reach 100 and still be inaccurate or never cited.
  • Plain text can score at most 80, because heading ids and JSON-LD can only be read from pasted HTML.
  • Each of the seven checks is scored separately, so a low total points at the specific property that is missing.
  • Google's own eligibility bar for AI features is that a page is indexed and eligible to show with a snippet.

What it takes

  • The rendered HTML of a page, for a full score
  • Or its plain text, for a rough one

More tools

All of them free, all of them browser-only.

Browse the tools

How to use it

  1. 1Get the rendered HTMLOpen the page, view source or copy the DOM, and paste it in. Plain text works but cannot score heading ids or structured data.
  2. 2Run the scoreEach check is scored separately, so a low total tells you which property is missing rather than just that something is wrong.
  3. 3Fix the opening answer firstIt carries 25 of the 100 points, the largest single weight in the score, because it is the passage built to be quoted whole.
  4. 4Re-run after publishingScore the live page rather than the draft, because a content management system can strip ids and reorder markup.

Google’s own documentation is the right place to start: it says there are no additional requirements to appear in AI Overviews or AI Mode, and no special optimizations are necessary. A page has to be indexed and eligible to show in Search with a snippet. This checker does not claim otherwise. What it scores is narrower and still useful: whether an article is built so that a single passage can be lifted, quoted and attributed without rewriting.

What this AEO checker scores

Seven checks, 100 points in total. The weights are this tool’s editorial judgment, not a Google formula.

CheckPointsPasses when
Opening answer of 40 to 60 words25The answer paragraph is 40 to 60 words (25 to 80 earns 12)
Question-shaped headings15At least one h2 or h3 reads as a question
Headings carry stable ids10Every h2 and h3 has an id attribute
Standalone list items15Three or more list items run 8 words or more
Outbound citations15Two or more absolute links (5 points each)
Structured data present10At least one JSON-LD script block exists
Publication date exposed10A time element with datetime, or datePublished

In HTML mode the answer paragraph is the element marked data-answer if there is one, then the first paragraph inside main or article, then the first paragraph on the page. Question headings are those starting with how, what, why, when, where, which, who, can, does, do, is, are or should, or ending in a question mark.

Google’s AI optimization guide asks for content that is well written and easy to follow, organized in paragraphs and sections with headings that give a clear structure. Its AI features page asks that structured data match the visible content. Those are the two documented ideas this score turns into countable checks.

How to read the results

The total is colored green at 80 or above, amber from 55 to 79 and red below 55. Each check shows PASS or WORK with its points and a one-line note, such as “First paragraph is 72 words” or “2 of 5 headings read as a question”. Fix the WORK rows in points order: the opening answer carries the most, then headings, lists and citations.

Two limits matter. In plain text mode, heading ids and structured data always score zero, so the ceiling is 80. And the citation check counts any absolute link, including links to your own site, so it measures link presence rather than source quality.

Worked example

Paste this as plain text: a 51-word opening paragraph defining a canonical tag, then two markdown headings, ”## What does a canonical tag do?” and ”## Setting it up”, three bullets (two full sentences and one two-word fragment), one Google documentation URL and the date 2026-09-25.

The tool returns 54, in red:

  • Opening answer: 25 of 25, “First paragraph is 51 words.”
  • Question-shaped headings: 8 of 15, “1 of 2 headings read as a question.”
  • Heading ids: 0 of 10, “Paste HTML to check this.”
  • Standalone list items: 6 of 15, two qualifying items at 3 points each.
  • Outbound citations: 5 of 15, one link, below the pass mark of two.
  • Structured data: 0 of 10, “No JSON-LD block found.”
  • Publication date: 10 of 10.

In plain text mode a heading only counts as a question if it ends in a question mark, because the markdown hashes sit at the start of the line. Adding a second source and a third full-sentence bullet would lift the score to 62 and turn both of those checks to PASS.

Common mistakes it catches

  • An opening that clears its throat. A 90-word scene-setting intro scores zero on the largest check, even when the real answer sits two paragraphs down.
  • Headings a CMS strips of ids. Drafts often have ids that the published template drops, which is why the live page is worth re-scoring.
  • Fragment bullets. Three-word list items read fine in context and mean nothing once quoted alone.
  • Structured data that does not exist on the live page. A plugin that fails silently shows up as “No JSON-LD block found.”
  • No parseable date. A date written only as “last Tuesday” gives nothing to read.

What it deliberately does not score is accuracy, authority or originality. Google’s guide puts the weight on unique, non-commodity content with a first-hand point of view, and no script can measure that.

Sources

  1. 1.AI features and your website - Google Search CentralPrimary
  2. 2.Google's guide to optimizing for generative AI features on Search - Google Search CentralPrimary

Common questions

Does a high score mean an answer engine will cite my page?

No, and any tool claiming otherwise is guessing. This scores whether a page is structurally easy to extract and attribute. Whether a system chooses your page over another depends on authority, relevance to the specific prompt and factors none of us can observe.

Does Google require special markup or an llms.txt file for AI Overviews?

No. Google's AI features documentation says there are no additional requirements or special optimizations needed, and its AI optimization guide says you do not need new machine-readable files, AI text files or special markup to appear in Google Search.

Why does it weight the opening paragraph so heavily?

It is worth 25 of the 100 points because a self-contained answer near the top is the passage easiest to lift whole. That weighting is this tool's editorial judgment, not a published Google rule, and the 40 to 60 word window is the same budget Surfil applies to its own articles.

Why does plain text score lower than HTML?

Heading ids and JSON-LD only exist in markup, so in plain text mode those two checks always score zero and the ceiling is 80. Paste the rendered HTML of the live page for the full score.

Is my content uploaded anywhere?

No. The page loads a small script that parses the text in your browser. There is no server involved, no request carrying your content, and nothing stored.