What does the citability score actually measure?
It combines five pattern-based sub-scores — directness, self-containedness, concreteness, sentence length, and passage length — into a single weighted 0-100 number. Each sub-score looks for a specific textual pattern associated with passages that read as standalone, quotable claims, such as opening with a concrete subject instead of a pronoun.
Is this a live scan of how AI systems cite my content?
No. This is a deterministic heuristic that runs entirely in your browser against the text you paste; it does not fetch a URL, call a model, or send anything to a server. It cannot know whether any specific AI system will actually cite a given passage, and a high score is not a guarantee of citation.
Why did my passage score low on self-containedness?
The self-containedness check flags passages whose first word is a pronoun that depends on unincluded context, such as "This," "It," "They," "These," or "That." Replacing that opening word with the actual subject usually raises the score, since the passage can then be read and understood on its own.
What counts toward the concreteness sub-score?
The tool counts numerals, percent and currency symbols, and capitalized proper-noun-like tokens per 100 words as a rough proxy for specific, quotable facts versus vague generalities. It's a surface-level pattern match, not true named-entity recognition, so unusual capitalization or acronyms can occasionally skew the count.