Paste a URL or raw text and get a scored breakdown across five AI-readability dimensions — plus a BLUF rewrite of your opening sentences, ready to copy in. Find out exactly why AI engines skip your content and what to change first.
The tool analyses your content against the five signals most correlated with AI citation rates. You can run it on any live page or paste content directly — useful for drafts, gated pages, or content you're editing before it goes live.
Traditional readability tools measure sentence length and syllable count — metrics built for human comprehension. AI readability is a different problem. A model extracting an answer from your page cares about answer location, answer density, and structural clarity. These five dimensions capture what actually predicts citation rate.
Rudolf Flesch developed his readability formula in 1948 to help the US Army write clearer field manuals. The formula counts syllables and sentence length — a reasonable proxy for how hard a human finds text to read. Copywriters and content teams have used variants of it ever since.
It tells you nothing useful about AI citation rates. A page written at a sixth-grade reading level with short sentences and simple words can still score an F on AI readability if it buries the answer in paragraph six. Conversely, a technically dense page with long sentences can score an A if it leads with a precise, direct answer and structures the rest as clearly labelled sections.
The difference comes down to what's being optimised. Flesch-Kincaid optimises for human comprehension — the experience of reading through a document linearly. AI readability optimises for machine extraction — the ability of a model to locate and quote a specific answer from anywhere in the document without reading the whole thing.
LLMs don't read your page from top to bottom the way a person does. They tokenise it, weight different sections by position and structure, and pull the highest-confidence answer chunk. A page that's structured for that process — BLUF opening, headed sections, FAQ block — performs well regardless of sentence complexity. A page that reads beautifully but puts its conclusions last performs poorly.
Content that scores well on traditional SEO metrics often performs poorly in AI search. This tool is for any team that's noticed the gap — pages ranking on page one that never appear in AI-generated answers.
Most pages score C or below not because the content is bad but because it's structured for a human reader scanning from top to bottom. Restructuring for AI extraction is usually faster than rewriting — you're moving and reformatting existing content, not generating new ideas.