# =============================================================
# Sushi Maru Winnipeg - robots.txt
# Canonical site: https://www.sushimaruwpg.ca
# =============================================================

# -------------------------------------------------------------
# Default policy for all crawlers
# -------------------------------------------------------------
User-agent: *
Allow: /

# Block legal pages from indexing
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html

# Block infrastructure / non-content paths (no-op if absent)
Disallow: /config/
Disallow: /account/
Disallow: /search/
Disallow: /api/
Allow: /api/ui-extensions/
Disallow: /static/

# Block common query-string variants (faceted/duplicate URLs)
Disallow: /*?*author=*
Disallow: /*?*tag=*
Disallow: /*?*month=*
Disallow: /*?*view=*
Disallow: /*?*format=*

# -------------------------------------------------------------
# AI / LLM crawlers
# llms.txt is intentionally crawlable so AI agents can read it,
# but it is served with `X-Robots-Tag: noindex` (see _headers)
# so it stays out of search engine indexes.
# -------------------------------------------------------------
User-agent: GPTBot
User-agent: ChatGPT-User
User-agent: OAI-SearchBot
User-agent: anthropic-ai
User-agent: Claude-Web
User-agent: ClaudeBot
User-agent: CCBot
User-agent: Google-Extended
User-agent: PerplexityBot
User-agent: cohere-ai
User-agent: FacebookBot
Allow: /
Allow: /llms.txt
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html

# -------------------------------------------------------------
# Google Ads crawlers (allow full access for ad quality checks)
# -------------------------------------------------------------
User-agent: AdsBot-Google
User-agent: AdsBot-Google-Mobile
User-agent: AdsBot-Google-Mobile-Apps
Allow: /

# -------------------------------------------------------------
# Heavy / aggressive crawlers - throttle
# -------------------------------------------------------------
User-agent: Baiduspider
Crawl-delay: 10

# -------------------------------------------------------------
# Sitemap
# -------------------------------------------------------------
Sitemap: https://www.sushimaruwpg.ca/sitemap.xml