# antevo.ch
#
# Deliberately permissive. Each product on this origin (trademark, wealth, mandate,
# executive) keeps whatever posture it sets in its own pages; this file does not
# decide that for them.

# THE DESKS. Both app paths are disallowed for EVERY crawler, and the rules are
# repeated in each group below because that is the only way to say it.
#
# Disallow binds to the most recent User-agent group, not to the file. These two
# lines used to sit only under the AI-crawler group further down, which meant
# `User-agent: *` carried `Allow: /` and nothing else — so Googlebot, Bingbot and
# every audit crawler were free to walk both desks while GPTBot and ClaudeBot were
# correctly kept out. The exact inverse of what the comment claimed. Found
# 2026-08-30 by parsing the served file into groups rather than reading it.
#
# No trailing slash, deliberately. These are prefix matches: `/trademark/app/` would
# have covered /trademark/app/mark but NOT /trademark/app itself, which is the
# sign-in shell and the one most likely to be linked.
#
# This keeps the desks out of the CRAWL. It does not guarantee they stay out of the
# INDEX: a disallowed URL that someone links to can still be listed URL-only,
# because the blocked crawler never gets to read a noindex. The real fix is
# `noindex` served by the apps themselves, at which point these Disallow lines
# should come out — a page can be crawled or it can be noindexed, not both.
User-agent: *
Allow: /
Disallow: /trademark/app
Disallow: /mandate/app

# AI crawlers and answer engines are welcome on the public pages.
#
# `User-agent: *` above already permits them — and for Google-Extended and
# Applebot-Extended, absence of a rule IS permission, since those are opt-OUT
# controls. Naming them anyway is a statement of intent: these pages exist to answer
# a question, and being quoted accurately by an assistant is the same job as ranking.
# It also means a future blanket Disallow has to disable this deliberately rather
# than by accident.
#
# Naming a group also COSTS something, which is the trap above: a named group takes
# the agent OUT of `*` entirely, so every rule that should apply to it has to be
# restated here. Nothing is inherited.
#
# See /llms.txt for what this site is, and /trademark/llms.txt for the detail.
User-agent: GPTBot
User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: ClaudeBot
User-agent: Claude-User
User-agent: Claude-SearchBot
User-agent: PerplexityBot
User-agent: Perplexity-User
User-agent: Google-Extended
User-agent: Applebot-Extended
User-agent: CCBot
User-agent: meta-externalagent
Allow: /
Disallow: /trademark/app
Disallow: /mandate/app

# The index above points at each product's own sitemap. One of them is not on this
# host: the Executive notes (/executive/notes/<date>/, 117 URLs on 2026-10-16) are
# listed by a sitemap the API serves at api.antevo.ch. Under the sitemaps protocol a
# sitemap may only list URLs of ANOTHER host if that host's robots.txt names it —
# so the reference has to live here, on antevo.ch, or engines may drop the whole
# file. Bing had indexed roughly 96 of the site's 218 sitemap URLs that day.
#
# The other three are listed too. An engine that reads the index finds them anyway;
# one that only reads robots.txt now does as well.
Sitemap: https://antevo.ch/sitemap.xml
Sitemap: https://api.antevo.ch/v1/public/executive/sitemap.xml
Sitemap: https://antevo.ch/trademark/sitemap.xml
Sitemap: https://antevo.ch/ask/sitemap.xml
Sitemap: https://antevo.ch/pages.xml