What does a machine see on this page?
One page, measured against the browser's own account of it: the share of the words a reader sees that survive extraction, what the page is worth before its scripts run, whether reading order was measured or assumed, what was laid out and hidden, and what structure reaches the other side. Every part carries the sentence it was computed from.
One page, one render, no crawl. Every request identifies itself as webgraph; the browser is a real Chromium under its own name. For a whole site — robots.txt per bot, walls, dead links, a suggested llms.txt — run a site report.