# Taj Mahal Food Robots.txt
# https://tajmahalfoodlon.ca/

# ----- All bots -----
User-agent: *
Allow: /
Disallow: /components/
Disallow: /llms.txt
Disallow: /update-menu.js
Disallow: /menu.js
Disallow: /*?*

# ----- Major AI / training bots -----
# Allow indexing of public pages but keep llms.txt private (it is for
# AI agents that fetch it directly, not for general search indexing).
User-agent: GPTBot
User-agent: ChatGPT-User
User-agent: CCBot
User-agent: anthropic-ai
User-agent: Claude-Web
User-agent: ClaudeBot
User-agent: Google-Extended
User-agent: PerplexityBot
User-agent: FacebookBot
User-agent: cohere-ai
Allow: /
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html
Disallow: /llms.txt

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

# ----- Heavy-load crawler throttling -----
User-agent: Baiduspider
Crawl-delay: 10

# ----- Sitemap -----
Sitemap: https://tajmahalfoodlon.ca/sitemap.xml