# robots.txt for W2 Films
# https://www.w2films.com.br/robots.txt
# Updated: 2025-01-16

# ============================================
# ALLOW ALL CRAWLERS (SEO + AI BOTS)
# ============================================

# Default: Allow all bots
User-agent: *
Allow: /
Crawl-delay: 1

# ============================================
# AI BOTS - EXPLICITLY ALLOWED
# ============================================

# OpenAI GPT Bots
User-agent: GPTBot
Allow: /
Crawl-delay: 1

User-agent: ChatGPT-User
Allow: /

# Anthropic Claude
User-agent: anthropic-ai
Allow: /
Crawl-delay: 1

User-agent: Claude-Web
Allow: /

# Common Crawl (used by many AI models)
User-agent: CCBot
Allow: /
Crawl-delay: 1

# Perplexity AI
User-agent: PerplexityBot
Allow: /
Crawl-delay: 1

# Google Extended (for AI training)
User-agent: Google-Extended
Allow: /

# Google Bard/Gemini
User-agent: Google-InspectionTool
Allow: /

User-agent: GoogleOther
Allow: /

# Bing AI / Copilot
User-agent: Bingbot
Allow: /

# Meta AI
User-agent: FacebookBot
Allow: /

User-agent: Meta-ExternalAgent
Allow: /

# Diffbot (AI data extraction)
User-agent: Diffbot
Allow: /

# Applebot (Siri, Spotlight)
User-agent: Applebot
Allow: /

# Bytedance/TikTok Bot
User-agent: Bytespider
Allow: /

# Yandex (Russian search/AI)
User-agent: YandexBot
Allow: /

# Baidu (Chinese search/AI)
User-agent: Baiduspider
Allow: /

# ============================================
# SEARCH ENGINE BOTS
# ============================================

User-agent: Googlebot
Allow: /

User-agent: Googlebot-Image
Allow: /

User-agent: Googlebot-Video
Allow: /

User-agent: Slurp
Allow: /

User-agent: DuckDuckBot
Allow: /

# ============================================
# DISALLOW PATTERNS
# ============================================

# Disallow only unnecessary/sensitive files
Disallow: /cgi-bin/
Disallow: /*.php$
Disallow: /*.inc$
Disallow: /admin/
Disallow: /private/
Disallow: /*.log$

# ============================================
# SITEMAPS
# ============================================

Sitemap: https://www.w2films.com.br/sitemap.xml
Sitemap: https://www.w2films.com.br/sitemap-images.xml
Sitemap: https://www.w2films.com.br/sitemap-videos.xml

# ============================================
# AI DISCOVERY FILES
# ============================================

# AI plugin manifest
Allow: /.well-known/ai-plugin.json
Allow: /ai-discovery.json
Allow: /openapi.json
Allow: /schema-advanced.json

# ============================================
# HOST
# ============================================

Host: https://www.w2films.com.br