User-agent: *
Allow: /

# The owner-facing funnel landings are NOT disallowed, deliberately. They
# used to be — their crawlable surface scales with the directory, so the
# rule was there to save the fetch their noindex cannot. But a disallowed
# URL is never fetched, so its noindex is never read, and Google indexes it
# from the anchor alone: on 22 Aug 2026 seven `/{locale}/signup?number=…
# &name=…` URLs surfaced as "Indexed, though blocked by robots.txt", with
# the business name in the query string. Letting the fetch through is what
# makes the noindex enforceable. rel="nofollow" on the links that mint them
# still holds discovery down — see components/card/CallerLookupCard/
# CardClaimStrip.js.

# Block AI *training* crawlers only. Search / assistant / user-initiated AI
# bots stay allowed under "*" (OAI-SearchBot, ChatGPT-User, Claude-User,
# Claude-SearchBot, PerplexityBot, Applebot, GoogleOther, bingbot, …).
User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: CCBot
Disallow: /

User-agent: Bytespider
Disallow: /

User-agent: meta-externalagent
Disallow: /

# Opt-out tokens (not fetchers): drop AI training without touching Google
# Search or Apple (Siri/Spotlight) indexing.
User-agent: Google-Extended
Disallow: /

User-agent: Applebot-Extended
Disallow: /

# AI assistants: Calmido is an MCP server with one read-only lookup tool at
# /api/mcp — see /nl-en/developers and /llms.txt.

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