zellahq.com

mx

route1.mx.cloudflare.net

route2.mx.cloudflare.net

route3.mx.cloudflare.net

ns

art.ns.cloudflare.com

donna.ns.cloudflare.com

caa

txt

google-site-verification=oowJDVR64tH6iiAc8GNxSTffyNdnJ-2qyITDWE_kblk

mailerlite-domain-verification=8527620a49b50d3af20cfac6df7ea736d6125361

v=spf1 include:_spf.mlsend.com include:_spf.mx.cloudflare.net include:_spf.google.com ~all

dmarc

# zellahq.com

User-agent: *
Allow: /
Disallow: /admin

# ── AI crawlers: explicitly welcome ──────────────────────────────────────────
# `Allow: /` above already permits these, but the tokens are listed explicitly
# so the intent survives any future tightening of the wildcard rule, and so
# Google-Extended / Applebot-Extended (which are OPT-OUT tokens with no crawler
# behind them) are clearly opted IN rather than left ambiguous.
# Tokens verified against the vendors' own docs, not a blog post.
#
# ONE GROUP, NOT TEN. Each of these used to be its own group carrying a lone
# `Allow: /`, and a robots.txt group is winner-takes-all: a crawler obeys the
# most specific group that names it and inherits NOTHING from `User-agent: *`.
# So naming these bots to be generous to them also, silently, handed all ten of
# them /admin - the one path the wildcard group exists to keep out. Consecutive
# user-agent lines followed by one rule block (RFC 9309 2.2.1) apply to every
# agent listed, so the tokens stay explicit and the rules stay in one place.

# OpenAI - developers.openai.com/api/docs/bots
User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: GPTBot
# Anthropic - support.claude.com
User-agent: ClaudeBot
User-agent: Claude-User
User-agent: Claude-SearchBot
# Google AI (Gemini / AI Overviews training token)
User-agent: Google-Extended
# Apple Intelligence
User-agent: Applebot-Extended
# Perplexity
User-agent: PerplexityBot
# Common Crawl (feeds many downstream models)
User-agent: CCBot
Allow: /
Disallow: /admin

Sitemap: https://zellahq.com/sitemap-index.xml