Collect page text

Fetch your live pages and report what actually arrives on the wire — status, title, meta description, how many words of prose survive with no JavaScript engine, headings, code blocks, links — beside the character count of the source we store for the same path. The row worth the call is the GAP between those two: a page that is 8 000 characters of markdown in the repository and 40 words on the wire renders client-side, which makes it perfect to every check that reads the source and unquotable to every assistant that reads the page. Nothing else on this server holds both ends of that comparison at once. Returns an evidence record and the exact fetch_url calls that produced every row, so you can re-run it by hand and get the same answer. No scores, no findings, no opinion — it is the raw material the audits are built on, at a sixtieth of the price. Use it for 'what does our page actually serve', 'is this page rendering server-side', 'check these three URLs', 'do our titles and descriptions exist', «что реально отдаёт страница», «рендерится ли она без JS», «проверь эти адреса». It does not say what the gap MEANS — for that, ask docsbook_expert with what you are trying to achieve and it names what to compare these rows against. collect_ai_citability is the neighbouring probe, scoring whether an answer engine can fetch and quote the same pages. Returns a validated collect_page_text.v1 payload: an evidence map, the normalised rows behind it, and a reproduce block naming the exact MCP calls and arguments that produced every row — run them yourself and you get the same answer. There is no model in the path, so there are no findings, no scores and nothing to disbelieve; interpretation is what the audits charge for. Changes nothing; safe on a read-only token. Evidence, not a verdict — nothing here says what the gap MEANS. If you have not already asked docsbook_expert how to read it, ask: it says what this evidence is worth against, and what to do with it. One call, changes nothing.