# Kai Ping Restaurant - robots.txt
# https://kaipingrestaurant.ca/

# ============================================================
# Default rules for all crawlers
# ============================================================
User-agent: *
Allow: /
Disallow: /components/
Disallow: /*?*author=*
Disallow: /*?*tag=*
Disallow: /*?*month=*
Disallow: /*?*view=*
Disallow: /*?*format=*

# Do not index the AI reference file
Noindex: /llms.txt
Disallow: /llms.txt

# ============================================================
# AI / LLM crawlers
# Allow site content and the llms.txt index, but keep legal
# pages out of AI training / retrieval corpora.
# ============================================================
User-agent: GPTBot
User-agent: ChatGPT-User
User-agent: OAI-SearchBot
User-agent: CCBot
User-agent: anthropic-ai
User-agent: Claude-Web
User-agent: ClaudeBot
User-agent: Google-Extended
User-agent: PerplexityBot
User-agent: cohere-ai
User-agent: FacebookBot
User-agent: Applebot-Extended
User-agent: Bytespider
Allow: /
Allow: /llms.txt
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html

# ============================================================
# Google Ads bots
# ============================================================
User-agent: AdsBot-Google
User-agent: AdsBot-Google-Mobile
User-agent: AdsBot-Google-Mobile-Apps
Allow: /

# ============================================================
# Heavy / low-value crawlers
# ============================================================
User-agent: Baiduspider
Crawl-delay: 10

# ============================================================
# Sitemap
# ============================================================
Sitemap: https://kaipingrestaurant.ca/sitemap.xml