Paste a URL or the full text of an article. Get a 0–100 score for how likely Perplexity, ChatGPT, Claude, Gemini, and Google AI Overviews are to cite it — across 14 signals.
Option A: paste a URL — we'll try to fetch the public HTML. Option B: paste the full HTML or article text — recommended for private pages or when fetch is blocked by CORS.
AI systems rarely cite very short posts. Aim for 800+ words.
Numbers and dates signal factual content. Target 15+ per 1000 words.
At least two quotations attributed to named humans indicates real reporting.
Three or more links to .gov, .edu, .org, Wikipedia, Wikidata, PubMed, arXiv, or established news domains.
At minimum, Article/BlogPosting with author Person and datePublished. Extra credit: Organization + Breadcrumb + FAQPage.
<link rel="canonical"> so AI systems don't index duplicates.
Named author with a byline or Person schema.
Visible dates (and in schema). AI overwhelmingly prefers recent content.
Phrases like "we surveyed", "our data shows", "we analyzed", "our team found" indicate first-party research AI models prefer.
HTML lists and tables are machine-extractable; AI answers reuse them directly.
Subheadings phrased as questions map directly to common prompts and increase citation probability.
Side-by-side comparisons are one of the highest-cited structural patterns.
Proper <article>, <h1>, <section>, <time> elements make extraction unambiguous.
A discoverable /llms.txt signals you welcome AI ingestion; absence is a soft penalty for large, AI-aware sites.
This audit uses on-page heuristics only. It does not query Perplexity, ChatGPT, Claude, or Gemini directly. Citation behavior depends on many factors outside any single page — domain authority, recency, prompt context, RAG pipeline setup. Use the score as a relative indicator of structural readiness, not a guaranteed citation forecast.