# robots.txt — Sistema Colmeia (Grupo SOITIC)
# Site institucional. Permitimos indexação ampla (incluindo bots de IA)
# para maximizar presença em buscas e em respostas geradas por LLMs.

User-agent: *
Disallow: /painel
Disallow: /painel/
Allow: /

# ---- Bots de busca tradicionais ----
User-agent: Googlebot
Allow: /

User-agent: Bingbot
Allow: /

User-agent: DuckDuckBot
Allow: /

User-agent: Applebot
Allow: /

# ---- Bots de IA (treino e/ou search-augmented generation) ----
# Permitidos por padrão pra ganhar visibilidade nas respostas dos LLMs.
# Pra bloquear treino mantendo busca, troque "Allow: /" por "Disallow: /".

# OpenAI (ChatGPT crawl + search)
User-agent: GPTBot
Allow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

# Anthropic (Claude)
User-agent: ClaudeBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: Claude-Web
Allow: /

# Google (treino do Gemini, separado do Googlebot de busca)
User-agent: Google-Extended
Allow: /

# Perplexity
User-agent: PerplexityBot
Allow: /

# Meta (treino e busca)
User-agent: meta-externalagent
Allow: /

User-agent: FacebookBot
Allow: /

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

# Cohere
User-agent: cohere-ai
Allow: /

# Common Crawl (alimenta vários LLMs open-source)
User-agent: CCBot
Allow: /

# ByteDance (TikTok)
User-agent: Bytespider
Allow: /

# Amazon
User-agent: Amazonbot
Allow: /

# ---- Sitemap ----
Sitemap: https://sistemacolmeia.com.br/sitemap.xml