# Crawl everything, and use it to help people: search it, and quote it to a
# reader through an assistant. Do not train on it. That is the owner's ruling
# (13 October 2026; the ruling of 6 September allowed training too) on the
# three uses the Content Signals Policy (contentsignals.org) lets a site rule
# on, and the signal line says it in the policy's own terms. The same ruling
# is a reservation of text-and-data-mining rights under the EU's copyright
# directive, in the form that law reads: /.well-known/tdmrep.json, and the
# tdm-reservation meta in every page's head, each naming the colophon's
# paragraph on the words as the policy. A rule at Cloudflare's edge refuses
# the training crawlers besides, though any crawler may read this file
# (BathLimited/cf-tf, bots.tf). Lighthouse's robots audit has known the
# directive since 13.0 (January 2026); the pinned @lhci/cli still runs 12.6,
# which calls it unknown, so CI skips that one audit (lighthouserc.json) and
# tests/robots.spec.ts holds this file instead, to a stricter rule.
User-agent: *
Content-Signal: search=yes, ai-input=yes, ai-train=no
Allow: /

Sitemap: https://bath.limited/sitemap-index.xml