Aeopro applies the self-containment answer-first (BLUF) test to each H2 on your page and gives you a traffic-light status: green if that paragraph can be quoted as is, red if the AI has nothing to grasp.
When an AI engine quotes you, it almost never quotes your entire page: it extracts the first paragraph that follows a heading. If that paragraph starts with 'nowadays' or 'this is very important', there’s nothing to extract.
Aeopro scans up to 25 H2 sections of the page and evaluates, one by one, the first paragraph that comes after the heading: that is the snippet an answering engine has on hand when it wants to quote you. Each section is marked green, amber, or red, and the reason is always specific, never generic advice.
The test is deterministic and does not consume any AI resources. It marks red the paragraph that has fewer than 12 words —there's no answer, just a transition— and amber for four very typical vices: starting with a dangling anaphora ('this', 'as we mentioned', 'wie gesagt') that forces reading the previous text; starting with filler ('nowadays', 'in today's world', 'de nos jours'); being too long to be a snippet; and not containing the term from the heading even once, making it impossible to pair with the question that would bring it.
The length thresholds adjust to the language of the page, which Aeopro detects automatically: German compresses more (85 reference words), English stretches (110), and Spanish, French, Italian, and Portuguese remain at 100, with the test only complaining starting from 140% of that figure. Thus, a page in German does not fail for writing as it is done in German. If more than 40% of your sections fail the test, the report also opens a finding with the three worst by name.
The AI layer is optional and upon request. The 'Refine with AI' button —with its cost in tokens displayed before clicking— sends up to 15 sections for meaning review, not form: it returns answers, partially or not answering, with a one-line reason in your language, and is cached with a date so you don't pay for it twice. And if a section turns out poorly, you can edit it and the AI will rewrite it in answer-first format, with a better heading if applicable and without inventing information that your text does not have.
Traffic-light status by section, not an overall score
you can quickly see which H2s are quotable and which are not, with the exact reason for each (short paragraph, anaphora, filler, too long, missing the heading term). You know what to rewrite and in what order.
Free to repeat
the test is part of the deterministic checks of the report and does not consume tokens. You rewrite a paragraph, reanalyze, and check if it turned green, as many times as you want.
Thresholds by language, detected alone
what is a long paragraph in German is not one in English. Aeopro detects the language of your page and adjusts the reference (85 / 110 / 100 words) before judging anything.
Catches the dangling anaphora, the mistake that nobody reviews
a first paragraph that starts with 'this', 'as we said', or 'in addition to the above' is perfect for those reading from top to bottom and junk for those extracting a loose fragment. The test detects it in six languages.
Detects filler
'nowadays', 'in today's world', 'in the modern world'. It's the most common way to waste the most valuable paragraph of the section, and it gets marked in amber with a name and surname.
Checks that the paragraph talks about the heading
if the first paragraph does not mention the H2 term even once, the engine cannot match it with the question that would bring it. It's a silly check to explain and surprisingly effective.
Sense judgment with AI, only if you ask for it
the deterministic test looks at the form; the AI button reads the paragraph and tells you if it truly answers the heading, with a one-line reason. Up to 15 sections, the cost written before pressing and the result cached with a date.
From the red section to the rewritten text
you touch the section and the AI returns the paragraph in BLUF format, with a better heading if needed and an explicit warning when the section lacks sufficient substance. It does not fill gaps with information that you did not write.

Paste a URL and Aeopro will return the answer-first traffic-light status for each section, with the reason for each amber and red. Without using AI.