# Bollywood Box Robots.txt
# https://www.bollywoodboxcalgary.com

# Main directives for all bots
User-agent: *
Allow: /
Disallow: /config/
Disallow: /components/
Disallow: /api/
Disallow: /*?*author=*
Disallow: /*?*tag=*
Disallow: /*?*month=*
Disallow: /*?*view=*
Disallow: /*?*format=*

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

# AI Bot specific directives - Allow AI crawlers to access llms.txt
User-agent: GPTBot
User-agent: ChatGPT-User
User-agent: CCBot
User-agent: anthropic-ai
User-agent: Claude-Web
User-agent: cohere-ai
User-agent: PerplexityBot
User-agent: YouBot
Allow: /
Allow: /llms.txt
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html

# Google Extended (AI training) - Allow access
User-agent: Google-Extended
Allow: /
Allow: /llms.txt
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html

# Facebook Bot
User-agent: FacebookBot
Allow: /
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html

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

# Search engine crawlers - Block llms.txt from indexing
User-agent: Googlebot
User-agent: Bingbot
User-agent: Slurp
User-agent: DuckDuckBot
User-agent: Yandex
Allow: /
Disallow: /privacy-policy.html
Disallow: /terms-of-use.html
# Note: llms.txt noindex is handled via X-Robots-Tag header

# Crawl delay for heavy bots
User-agent: Baiduspider
Crawl-delay: 10

User-agent: SemrushBot
Crawl-delay: 5

User-agent: AhrefsBot
Crawl-delay: 5

# Sitemap location
Sitemap: https://www.bollywoodboxcalgary.com/sitemap.xml