# postnext.io
#
# ONE User-agent group, deliberately. A named group (GPTBot, Bytespider,
# Meta-ExternalAgent, ...) REPLACES the wildcard rules for that agent rather
# than adding to them, so giving a named crawler the same rules as `*` buys
# nothing today and silently excludes it from every rule later added to `*`.
# Requests to add a named AI-crawler group have been declined for this reason;
# written down here so the decision is not re-litigated. Asserted by
# scripts/seo/qa-all.js.
#
# LLMs: https://postnext.io/llms.txt
# (a comment, because no standard directive exists for this and compliant
# parsers ignore comments. It costs nothing and is discoverable by a human or
# an agent reading the file.)

User-agent: *

# Application surfaces. Never useful in an index, and crawling them spends
# budget the ~1,300-page public corpus needs.
Disallow: /api/
Disallow: /superuser/
Disallow: /admin/
Disallow: /account/
Disallow: /dashboard/
Disallow: /onboarding/
Disallow: /wp-admin/
Disallow: /wp-includes/

# Tracking parameters mint an unbounded set of duplicate URLs for any page that
# carries one. Only parameters that add nothing a crawler needs are listed:
# every public page is self-canonical, and a canonical tag handles duplicates
# better than a Disallow, which would stop Google seeing the canonical at all.
Disallow: /*?utm_source=
Disallow: /*?utm_medium=
Disallow: /*?utm_campaign=

# Signup and upgrade CTAs carry ?plan= and ?source= for attribution. The
# destination is not indexable in any variant.
Disallow: /signup?
Disallow: /upgrade?

# The blog's JSON index (repo postnext-blog). It feeds the /latest-blog-posts widget
# server-side; it is not a page and should not be crawled.
Disallow: /blog/search.json

Sitemap: https://postnext.io/sitemap.xml