Paste a page URL to see whether its content is structured for AI engines to extract and cite. The checker reviews self-contained paragraphs, direct answers, Q&A structure, step-by-step instructions, tables, and heading hierarchy. Each issue includes the relevant passage. Results are reported by signal, not as an opaque 0–100 score.
Free, no account required. Run 10 checks per day, or sign in for 30. Sign in with Google
The checker fetches the raw HTML once and does not run JavaScript, matching what a non-rendering AI crawler can access. It evaluates six signals automatically. The seventh, whether a sentence is genuinely quotable, requires human review.
Page analyzed
Little readable content found in the raw HTML
The seven signals
How to read the verdicts
✅ The checker found the expected structure.
⚠️ The structure passed the automated check or is borderline, but the content still needs manual review. Signal 2 always receives this verdict.
❌ The checker found a structural issue and shows the relevant evidence.
N/A This signal does not apply to this type of page. Do not add an FAQ, procedure, or table only to satisfy the check.
1 · Self-contained chunk
—
When quoted on its own, does each paragraph still make sense?
How this is measured
The checker measures how many substantial paragraphs begin with a context-dependent word such as "This", "It", or "They". It also finds phrases that rely on surrounding context, such as "see above" or "the table below". Each affected passage is quoted. Definition · How to fix it
2 · Direct-answer / TL;DR block
—
Is the answer in the first one or two sentences?
How this is measured
The checker looks for a concise opening paragraph before the first H2 and checks whether sections begin with supporting text rather than another heading. It cannot determine whether the opening actually answers the question, so the best possible verdict for Signal 2 is ⚠️. Definition · How to fix it
3 · Q&A / FAQ structure
—
Do headings phrased as questions have answers beneath them?
How this is measured
The checker finds headings phrased as questions from H2 through H4 and checks whether each one has content beneath it. It cannot determine whether the questions reflect real search demand. A page with no question headings receives N/A; do not add an FAQ unless the content calls for one. Definition · How to fix it
4 · Step-by-step structure
—
If the page explains a process, are the steps presented as a numbered list?
How this is measured
The checker looks for ordered lists and for procedure-style headings, such as "How to…", that are not followed by numbered steps. It also checks average step length. It does not assess whether each step begins with an imperative verb. Pages without a procedure receive N/A. Definition · How to fix it
5 · Citable tables
—
Can each row be understood on its own, with a clear caption and column labels?
How this is measured
The checker looks for a caption, or a heading directly above the table, and header cells. It quotes rows that lose their meaning when read on their own. Pages without data tables receive N/A. Definition · How to fix it
6 · Heading hierarchy
—
Does the page have one H1, properly nested headings, and no skipped levels?
How this is measured
The checker counts H1s and finds skipped levels, such as H2 followed by H4, and empty headings used for styling. This is the only signal that can be evaluated entirely from structure. When a level is skipped, the report shows both headings. Definition · How to fix it
7 · Quotable sentence
Manual review
Does each section contain one concise, standalone claim that still makes sense when quoted?
This signal requires human review. A structural check cannot determine whether a sentence makes a clear, useful claim. The playbook explains how to assess it, with examples. Definition
Why we do not give a score
Citability scores are verifiable only when their formulas are public. The tools reviewed in the citability audit playbook use undisclosed formulas, so their composite scores cannot be independently checked.
Instead, this checker reports each signal separately and quotes the evidence behind every issue, such as a paragraph that begins with a context-dependent pronoun, an H2 followed by an H4, or a table without headers. You can compare each finding directly with your page.
How the check works
The checker fetches the page once as raw HTML and does not run JavaScript. This matches what a non-rendering AI crawler can access. If the HTML contains almost no readable text, the checker reports that issue first and does not evaluate the six structural signals.
It extracts the main content from the main or article element. If neither is available, it uses the body after removing navigation, headers, footers, and sidebars. It then collects headings, paragraphs, ordered lists, and tables in document order.
It evaluates the six automated signals in the order shown above. The wording heuristics for pronouns, question headings, and procedural cues are calibrated for English and Chinese.
It reports one of four verdicts for each signal: ✅, ⚠️, ❌, or N/A. Every issue includes the relevant passage. Conditional signals (3, 4, and 5) return N/A when they do not apply to the page.
What this check can and cannot judge
This checker evaluates structure, not accuracy or authority. A page can pass all six automated signals and still go uncited. Retrieval, freshness, and E-E-A-T are separate factors; citability applies only after the page has been retrieved for grounding.
Signal 2 checks only whether a concise opening paragraph is in the expected position. It cannot determine whether that paragraph actually answers the question, so its best possible verdict is ⚠️.
Signal 7 is intentionally excluded from the automated verdicts. A regular expression can measure surface features such as sentence length, but it cannot determine whether a claim is genuinely quotable.
The wording heuristics are calibrated for English and Chinese. For other languages, the checker evaluates only language-independent structural signals and identifies anything it skips.
The checker does not run JavaScript. If content appears only after client-side rendering, it reports the raw page as having little or no readable content. For these pages, making the content available in the initial HTML is the first issue to address (see server-side rendering).
Frequently asked questions
Why does this tool not provide a 0–100 citability score?
A composite score can be independently verified only when its formula is public. The tools reviewed in §8 of the citability audit playbook use undisclosed weights, so their scores cannot be reproduced. This checker instead reports each signal separately and quotes the evidence behind every issue. You can verify each finding on your page.
What does N/A mean? Is a missing FAQ hurting my citability?
N/A means the signal does not apply to this type of page. Signals 3, 4, and 5 are conditional: a definition page does not need a step-by-step list, and a narrative page may not need an FAQ. Adding these structures where they do not belong is a fake fix. Address ❌ verdicts and leave N/A results alone.
Why does Signal 2 never show a green check?
The automated check can confirm only that a concise introduction appears before the first H2. It cannot tell whether the introduction actually answers the question, so a green check would overstate what the tool measured. Treat ⚠️ as a prompt to read the opening and ask whether the first sentence gives the main answer.
Why fetch without executing JavaScript?
JavaScript execution varies by crawler. This checker intentionally analyzes the raw HTML response so you can see what a non-rendering crawler can access. If content appears only after client-side rendering, the report may find little or no readable text. A browser-based checker could miss this issue.
If all six signals pass, will AI engines cite my page?
Not necessarily. This checker measures passage structure after retrieval. Indexing, crawler access, relevance, freshness, and source authority all affect whether an engine retrieves and trusts a page. Passing all six signals means the page's structure is less likely to be the bottleneck. The citability entry explains how these factors fit together.