Citability: can an answer engine quote you, and how easily

Before it matters how well a page is written, three things decide whether an answer engine may use it at all. A snippet directive can forbid any preview, so there is nothing to quote. robots.txt can refuse the search and answer crawlers outright. And the HTML the server sends can arrive nearly empty, with the page only appearing once a browser has run its scripts, which a fetcher may never do.

A closed gate is a verdict of Blocked, whatever else the page does well. A site’s verdict is the worst verdict of any of its pages, never an average, because one blocked page is one answer you cannot be the source of.

The gates are part of the core reading and run on every plan, including the free one. The public scan on the Citability page checks two of the three on any address without an account.

Past the gates, the score is a weighted sum of sections, each a question the page either answers or does not. Only sections that could be measured count, so a check that needs Search Console does not drag a site down for lacking the connection.

• Gates (40): snippet directives, crawler access, whether the served HTML carries the page.
• Rendering (25): how much of the page exists before scripts run.
• Crawler access (15): the edge, the firewall, per-agent refusals.
• Structure (12): headings that say what follows, answers close to the question.
• Substance (8): figures, dates, attribution, something worth quoting.
• Structured data (5): markup that agrees with the visible text.
• Media and links (5): alt text and outbound sources.

The site screen is built in the order a reader asks: can these pages be quoted, how well, what is the number made of, and which checks failed. Every check is listed worst first, as Failed, Warning, Passed or Not measured, with the ones that close a gate marked as such.

• Read it again: a single check re-run on the live pages, so a fix can be confirmed without waiting for the next reading.
• Draft the fix: a model writes the change, code checks it. Every number in a draft must be printed on the page or carried in its address. Text that can be rewritten gets a before and an after; everything else gets steps. When nothing survives the checks, the draft says so instead of pretending.
• What could not be measured: named, so an absent Search Console connection reads as absent rather than as a fault.
• Over time: the trend across readings, as far back as the plan keeps history.
• The card on the reading view: two sentences above Search Console, with a link to the breakdown.
• Citability across your sites, on the Agency plan: a portfolio sorted worst first, where blocked pages decide the week and only then the score.

Citability reads every page of every reading. There is no toggle and no upload. The module, from the Business plan, unlocks the score, the section breakdown, the trend, the check list with its drafts and re-checks, and fan-out.

Fan-out asks what an assistant would search for next after your page, and checks whether the page covers those follow-up questions. It is the one part with a cost outside the reading, so each plan carries a budget of follow-up queries per reading, with no counter to watch and no line on the invoice.

On plans without the module, the findings are still counted. The reading says how many citability signals it found and in how many kinds of check, and keeps the names to itself.

Answer engines publish very little about how they read, so the module stays inside what can be measured on your own pages and says so where it cannot.

• No answer engine publishes whether its fetcher runs JavaScript, so the rendering checks measure what the server returns, not what an engine may or may not do with it.
• The edge probe is one request per user agent, once per reading. It catches a firewall rule aimed at a named bot, but a single request is a hint, not proof, unless the server logs agree.
• It does not monitor what AI engines say about your brand, generate an llms.txt, serve separate Markdown versions for bots, or advise chunking content into micro-paragraphs.
• Structured data is checked for agreeing with the visible text. It is not called a citation lever, because there is no evidence that it is one.