# robots.txt - DataPulseX Enterprise SEO Configuration
# Last Updated: 2026-03-25
# Standard: Compliant with robots.txt protocol v1.0
# Structure: public_html/ (Hostinger)

# ===========================================================
# GENERAL RULES FOR ALL BOTS
# ===========================================================
User-agent: *

# ALLOW PUBLIC CONTENT
Allow: /
Allow: /index.php
Allow: /assets/
Allow: /sitemap.xml
Allow: /sitemap-*.xml

# PUBLIC PAGES
Allow: /dash/login.php
Allow: /dash/register.php
Allow: /dash/termos.php
Allow: /dash/privacidade.php
Allow: /dash/recuperar_senha.php
Allow: /dash/manual.php

# ===========================================================
# DISALLOW SENSITIVE AREAS
# ===========================================================

# Admin & Backend
Disallow: /dash/admin_*
Disallow: /dash/backend/
Disallow: /dash/api/
Disallow: /dash/dashboard.php
Disallow: /dash/analytics.js
Disallow: /dash/pixel.js

# Configuration & Database
Disallow: /*.env
Disallow: /*.env.example
Disallow: /*.sql
Disallow: /dash/.env
Disallow: /dash/.env.example
Disallow: /dash/*.sql
Disallow: /*/config.php
Disallow: /*/db.php
Disallow: /*/auth.php
Disallow: /*/mailer.php

# Private User Areas
Disallow: /dash/dashboard.php
Disallow: /dash/perfil.php
Disallow: /dash/meu_plano.php
Disallow: /dash/checkout.php
Disallow: /dash/logout.php

# Session Parameters
Disallow: /*?sessionid=
Disallow: /*?session_id=
Disallow: /*?uid=
Disallow: /*?user_id=
Disallow: /*?token=
Disallow: /*?auth=

# Marketing Parameters (Preserve for Analytics)
Allow: /*?utm_
Allow: /*?fbclid=
Allow: /*?gclid=
Allow: /*?msclkid=

# Database & System Files
Disallow: /backup/
Disallow: /temp/
Disallow: /cache/
Disallow: /logs/
Disallow: /test/
Disallow: /tests/

# ===========================================================
# BOT-SPECIFIC RULES
# ===========================================================

# Google Bot
User-agent: Googlebot
Allow: /
Allow: /assets/
Disallow: /dash/admin_*
Disallow: /dash/api/
Disallow: /dash/backend/
Crawl-delay: 0

# Bing Bot
User-agent: Bingbot
Allow: /
Disallow: /dash/admin_*
Disallow: /dash/api/
Crawl-delay: 1

# Yahoo Bot
User-agent: Slurp
Allow: /
Disallow: /dash/admin_*
Crawl-delay: 2

# DuckDuckGo Bot
User-agent: DuckDuckBot
Allow: /
Disallow: /dash/admin_*

# ===========================================================
# BLOCK AGGRESSIVE/BAD BOTS
# ===========================================================

User-agent: AhrefsBot
Disallow: /

User-agent: SemrushBot
Disallow: /

User-agent: DotBot
Disallow: /

User-agent: MJ12bot
Disallow: /

User-agent: DomainCrawler
Disallow: /

# ===========================================================
# CRAWL DELAY & REQUEST RATE
# ===========================================================

# Default crawl delay for all bots (milliseconds)
Crawl-delay: 1

# Request rate (requests per second) - optional
# Request-rate: 1/10

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

# Primary sitemap
Sitemap: https://datapulsex.com.br/sitemap.xml

# Language-specific sitemaps
Sitemap: https://datapulsex.com.br/sitemap-pt-br.xml
Sitemap: https://datapulsex.com.br/sitemap-en-us.xml
Sitemap: https://datapulsex.com.br/sitemap-es.xml

# Mobile sitemap
Sitemap: https://datapulsex.com.br/sitemap-mobile.xml

# News sitemap (if applicable)
# Sitemap: https://datapulsex.com.br/sitemap-news.xml

# ===========================================================
# SEARCH ENGINE VERIFICATIONS
# ===========================================================

# Google Search Console
# verify=abc123def456...

# Bing Webmaster Tools
# BingSiteAuth: BingSearchConsoleCode

# Yandex Webmaster
# yandex-verification: abcd1234efgh5678

# ===========================================================
# ADDITIONAL NOTES
# ===========================================================
# 
# This robots.txt implements:
# - Enterprise-grade crawl control
# - Security by obscurity for sensitive areas
# - Bot reputation management
# - Sitemap declarations for multiple languages
# - Crawl optimization for search engines
#
# For JSON version, visit: /robots.json
# For SEO audit, visit: /seo-audit.php
#