# ============================================
# robots.txt - Compulaser
# https://compulaserg.com.br/robots.txt
# Atualizado: 2026-07-14
# ============================================

# ============================================
# REGRAS GLOBAIS
# ============================================

User-agent: *
Disallow: /config.php
Disallow: /*config.php
Disallow: /.env
Disallow: /*.env
Disallow: /includes/
Disallow: /admin/
Disallow: /vendor/
Disallow: /cache/
Disallow: /logs/
Disallow: /tmp/

# Parâmetros de tracking e sessão
Disallow: /*?utm_*
Disallow: /*?fbclid=*
Disallow: /*?gclid=*
Disallow: /*?utm_source=*
Disallow: /*?utm_medium=*
Disallow: /*?utm_campaign=*
Disallow: /*?utm_content=*
Disallow: /*?utm_term=*
Disallow: /*?sessionid=*
Disallow: /*?PHPSESSID=*
Disallow: /*?ref=*

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

Sitemap: https://compulaserg.com.br/sitemap.xml

# ============================================
# BOTS DE BUSCA / RETRIEVAL (LIBERADOS)
# Geram citações e tráfego de volta para o site
# ============================================

# OpenAI Search (ChatGPT com busca web)
User-agent: OAI-SearchBot
Allow: /

# ChatGPT User Agent (navegação em tempo real)
User-agent: ChatGPT-User
Allow: /

# Perplexity Search
User-agent: PerplexityBot
Allow: /

# Claude Search (busca pública do Anthropic)
User-agent: Claude-SearchBot
Allow: /

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

# Google (Search + AI Overviews)
User-agent: Googlebot
Allow: /

User-agent: Google-Extended
Allow: /

# Apple Intelligence Search
User-agent: Applebot-Extended
Allow: /

# ============================================
# BOTS DE TREINAMENTO (BLOQUEADOS)
# Impedem que seu conteúdo alimente modelos de IA
# ============================================

# OpenAI GPTBot (treinamento)
User-agent: GPTBot
Disallow: /

# Anthropic Claude (treinamento)
User-agent: Claude-Web
Disallow: /

User-agent: ClaudeBot
Disallow: /

# Common Crawl (usado por muitos LLMs)
User-agent: CCBot
Disallow: /

# Meta AI / Llama (treinamento)
User-agent: Meta-ExternalAgent
Disallow: /

# ByteDance / TikTok AI (treinamento)
User-agent: Bytespider
Disallow: /

# Amazon Alexa AI (treinamento)
User-agent: Amazonbot
Disallow: /

# ============================================
# Crawlers de SEO / Auditoria
# ============================================

User-agent: AhrefsBot
Allow: /

User-agent: AhrefsSiteAudit
Allow: /

User-agent: SemrushBot
Allow: /

User-agent: SemrushBot-SA
Allow: /

User-agent: MJ12bot
Allow: /
Crawl-delay: 5

User-agent: DotBot
Allow: /
Crawl-delay: 10

User-agent: DataForSeoBot
Allow: /

# ============================================
# Crawlers de mídia social
# ============================================

User-agent: facebookexternalhit
Allow: /

User-agent: LinkedInBot
Allow: /

User-agent: Twitterbot
Allow: /

User-agent: WhatsApp
Allow: /

# ============================================
# Bloquear crawlers maliciosos / indesejados
# ============================================

User-agent: PetalBot
Disallow: /

User-agent: Barkrowler
Disallow: /

User-agent: BLEXBot
Disallow: /