# Auralis LLC — www.auralisllc.com
#
# Open to crawlers, including AI crawlers. Auralis is a small firm whose visibility
# depends on being found and cited; blocking assistant traffic would cost more than it
# protects. Everything here is public marketing copy.
#
# If that ever changes, add specific User-agent blocks rather than a blanket Disallow.
#
# ── CLOUDFLARE OVERRIDES THIS FILE IF YOU LET IT (found and disabled 2026-08-05) ──
#
# Cloudflare has a zone setting - Signals -> "Managed robots.txt" - which PREPENDS its own
# block to whatever is here, adding `Disallow: /` for ClaudeBot, GPTBot, CCBot,
# Google-Extended, Amazonbot, Applebot-Extended, Bytespider and meta-externalagent, plus a
# `Content-Signal: ai-train=no` line. It switched itself ON when the zone activated at
# cutover, and it applied zone-wide - even to autodiscover.auralisllc.com, which is an
# email hostname and not a website at all.
#
# For about half a day the served file therefore said the opposite of this one, and nothing
# in the repository could have told you: the file here was unchanged and correct, and the
# edge was serving something else. It was found by fetching the live URL, not by reading
# code. **If this file and the served file ever disagree again, check that toggle first.**
#
# Jen turned it off on 2026-08-05 (docs/CUTOVER-RUNBOOK.md §22), so the served file is once
# more byte-identical to this one - verified. Two notes for whoever revisits it:
#
#  - What it blocked was TRAINING crawlers, not assistants. Live AI-search agents
#    (PerplexityBot, OAI-SearchBot, ChatGPT-User, Claude's search agents) were never in its
#    list and were always allowed. So the argument above - that blocking assistant traffic
#    costs more than it protects - was not actually the argument against it.
#  - The real trade was small: whether Auralis content may enter training corpora and
#    Common Crawl. Jen's call was to allow it, for the chance a model knows the firm exists
#    without being sent to look. Blocking would have cost nothing in Google Search either
#    way, since Googlebot was never blocked.
#
# It is a dashboard toggle, so it can be re-enabled by a Cloudflare default change as
# easily as by a person. Re-check it if AI referral traffic ever falls off a cliff.

User-agent: *
Allow: /

Sitemap: https://www.auralisllc.com/sitemap-index.xml