# Stippi -- https://stippi.com/
# A curated index of professional conferences.
#
# Position, decided Sep 2026. Open to everything, minus the two crawlers whose
# only job is AI training.
#
# The reasoning. Stippi's value is being the record people cite, and organisers
# are the likeliest source of revenue -- they are paying for reach. A model that
# has never seen Stippi is less likely to name it even when it can fetch it, so
# withholding the index costs more than it protects. And robots.txt is a
# request, not a control: it is honoured by the well-behaved, who are also the
# ones most likely to attribute. Declining them buys protection from the least
# threatening actors and none at all from a determined scraper.
#
# Google-Extended and Applebot-Extended are the exception. They govern AI
# training only and have no effect on Google Search or Siri results, so
# declining them costs nothing at all.
#
# What actually protects the work is elsewhere: every record carries the date
# its programme was read, so a copy is provably a snapshot; and the derived
# layer -- 4,604 people resolved to distinct identities, a 149-term vocabulary --
# cannot be extended by anyone who did not do the deriving.
#
# Reversibility. Blocking later is worse than blocking now, because anything
# ingested stays ingested. Two things would change this call: the paid speaker
# plan becoming real revenue, or a copy turning up in the wild.

# Governs AI training only, not search ranking. Declined.
User-agent: Google-Extended
Disallow: /

User-agent: Applebot-Extended
Disallow: /

# Everyone else, including search and retrieval crawlers.
User-agent: *
Allow: /

Sitemap: https://stippi.com/sitemap.xml