User-agent: *
Allow: /
Allow: /llms.txt
Allow: /llms-full.txt

# Block admin / provider dashboard / private user account pages from indexing.
# These ALSO carry X-Robots-Tag: noindex, nofollow HTTP headers (netlify.toml)
# The header is the authoritative signal; this Disallow is an advisory hint.
Disallow: /admin
Disallow: /labs/
Disallow: /provider-
Disallow: /auth/
Disallow: /account
Disallow: /profile
Disallow: /settings
Disallow: /my-passes
Disallow: /my-bookings
Disallow: /my-subscription
Disallow: /my-subscriptions
Disallow: /my-ticket
Disallow: /ticket/
Disallow: /remove
Disallow: /vendor-dashboard

# Sitemap location. Both addresses serve the same index; /sitemap-index.xml was added
# Oct 15 2026 as a fresh address for Search Console (see netlify/edge-functions/sitemap.js).
Sitemap: https://madeira.fun/sitemap.xml
Sitemap: https://madeira.fun/sitemap-index.xml

# LLM/AI agent information
# See /llms.txt for a concise site overview
# See /llms-full.txt for detailed content structure and Schema.org types
# Language guide pages have Markdown versions: append .md to a /{locale}/{slug} URL

# Open data (free, no key, CC BY 4.0). Documentation, methodology and rule changes:
#   https://madeira.fun/hiking/permit-data#open-data
# Feeds (please cache; they are data, not pages, so they are not in the sitemaps):
#   /api/public/hiking-availability (+ .csv)             every bookable permit trail
#   /api/public/hiking-availability/{seo_slug} (+ .csv)  one trail, up to 120 days ahead
#   /api/public/madeira-weather (+ .csv)                 live weather by place (method: /weather/data)
# OpenAPI 3.1: /api/public/openapi.json
# Read-only MCP server: https://madeira.fun/api/mcp

# AI agents: ALL crawlers and live user-fetch agents (GPTBot, ChatGPT-User,
# OAI-SearchBot, ClaudeBot, Claude-User, Claude-SearchBot, PerplexityBot,
# Perplexity-User, Applebot, Meta-ExternalAgent, Bytespider, etc.) are welcome
# under the wildcard group above - max-visibility policy (May 2026 SEO
# decision). Deliberately NO dedicated per-crawler groups: under RFC 9309 a
# crawler that matches a dedicated group IGNORES the * group, so the old
# "User-agent: GPTBot / Allow: /" courtesy blocks accidentally waived the
# privacy Disallows above for exactly those bots (fixed Aug 2026). The two
# groups below are the exception: they are AI-usage POLICY TOKENS, not
# fetching crawlers - a dedicated group is the only way to signal them.

# Google AI grounding/training (Gemini, Vertex AI). Does NOT affect Google
# Search ranking, controls AI surface usage only.
User-agent: Google-Extended
Allow: /

# Apple Intelligence training opt-in (Applebot itself crawls under *).
User-agent: Applebot-Extended
Allow: /