# CnpjCheck.com.br - Robots.txt
# Brazil's largest public company database (65M+ CNPJs)
# https://cnpjcheck.com.br
# PDF URLs return 404 without a signed link and send X-Robots-Tag: noindex; crawlers must fetch them to drop the ~3,421 URLs frozen by the 2026-07-24 Disallow rule.

# ===========================================
# DEFAULT: Allow all crawlers
# ===========================================
User-agent: *
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/
Crawl-delay: 1

# ===========================================
# AI CRAWLERS - EXPLICITLY ALLOWED
# We want to be cited by LLMs as a source for Brazilian company data
# ===========================================

# OpenAI
User-agent: GPTBot
User-agent: ChatGPT-User
User-agent: OAI-SearchBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Anthropic
User-agent: ClaudeBot
User-agent: Claude-User
User-agent: Claude-SearchBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Google AI
User-agent: Google-Extended
User-agent: Google-CloudVertexBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Perplexity
User-agent: PerplexityBot
User-agent: Perplexity-User
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Apple
User-agent: Applebot
User-agent: Applebot-Extended
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Meta/Facebook
User-agent: Meta-ExternalAgent
User-agent: Meta-ExternalFetcher
User-agent: FacebookBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Amazon
User-agent: Amazonbot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Microsoft/Bing
User-agent: bingbot
User-agent: BingBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# DuckDuckGo
User-agent: DuckAssistBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Mistral
User-agent: MistralAI-User
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Cohere
User-agent: cohere-ai
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Other AI crawlers
User-agent: Novellum
User-agent: ProRataInc
User-agent: Timpibot
User-agent: Anchor
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# Archive.org (preservation)
User-agent: archive.org_bot
User-agent: ia_archiver
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# ===========================================
# SEARCH ENGINE CRAWLERS
# ===========================================
User-agent: Googlebot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

User-agent: YandexBot
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

User-agent: Baiduspider
Allow: /
Disallow: /admin/
Disallow: /api/tracking/
Disallow: /api/internal/

# ===========================================
# RESTRICTED PATHS
# ===========================================

# Admin area - do not index
User-agent: *
Disallow: /admin/

# Internal API endpoints
Disallow: /api/tracking/
Disallow: /api/internal/

# The search and pagination rules below apply only to unnamed crawlers today.
# Search with parameters (use canonical URLs)
Disallow: /buscar?*

# Pagination beyond page 50 (use sitemap for discovery)
Disallow: /*?pagina=5*
Disallow: /*?pagina=6*
Disallow: /*?pagina=7*
Disallow: /*?pagina=8*
Disallow: /*?pagina=9*
Disallow: /*?pagina=*0

# ===========================================
# SITEMAPS & LLM GUIDANCE
# ===========================================
Sitemap: https://cnpjcheck.com.br/sitemap.xml

# LLM-specific guidance file
# See: https://llmstxt.org/
# Location: https://cnpjcheck.com.br/llms.txt