# AI crawlers are deliberately allowed by the wildcard rule below.
#
# Before you "block AI bots", know that each vendor now ships separate bots for
# TRAINING and for LIVE RETRIEVAL, and only the retrieval ones decide whether we
# can be cited in an answer:
#
#   training  — GPTBot, ClaudeBot, Google-Extended, CCBot, Applebot-Extended
#   retrieval — OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User,
#               PerplexityBot, Perplexity-User, Googlebot (this is what feeds
#               AI Overviews, NOT Google-Extended)
#
# Blocking a retrieval bot removes us from that assistant's answers silently:
# no error, no report, just absence. AI answers already describe our
# infrastructure and pricing accurately and unprompted, so that would be a real
# loss. Block training bots if we ever want to; leave retrieval alone.

User-agent: *
Allow: /
Disallow: /auth
Disallow: /login
Disallow: /register
Disallow: /draft/
Disallow: /projects
Disallow: /deployments
Disallow: /billing
Disallow: /settings
Disallow: /admin

Sitemap: https://deploybase.eu/sitemap.xml