# 123B - Robots.txt - Tối ưu cho SEO & AI Crawler 2026
# https://www.123bpr.online/robots.txt

# ============================================
# Default: Allow all crawlers
# ============================================
User-agent: *
Disallow: /wp-admin/
Disallow: /wp-includes/
Disallow: /wp-content/plugins/
Disallow: /wp-content/cache/
Disallow: /wp-content/litespeed/
Disallow: /xmlrpc.php
Disallow: /wp-json/
Disallow: /readme.html
Disallow: /license.txt
Disallow: /wp-config.php
Disallow: /cgi-bin/
Disallow: /feed/
Disallow: /trackback/
Disallow: /?s=
Disallow: /search/
Disallow: /author/
Disallow: /comments/feed/
Allow: /wp-admin/admin-ajax.php
Allow: /wp-content/uploads/
Allow: /wp-content/themes/

# ============================================
# AI Crawlers - Selective Allow (cho Citation 2026)
# Cho phép AI crawler truy cập để được trích dẫn bởi AI Overviews
# ============================================

# GPTBot (OpenAI) - Allow for citation
User-agent: GPTBot
Allow: /
Disallow: /wp-admin/
Disallow: /wp-includes/

# ChatGPT-User (OpenAI real-time browsing) - Allow
User-agent: ChatGPT-User
Allow: /

# ClaudeBot (Anthropic) - Allow for citation
User-agent: ClaudeBot
Allow: /
Disallow: /wp-admin/

# Claude-Web (Anthropic real-time)
User-agent: Claude-Web
Allow: /

# PerplexityBot - Allow for search index
User-agent: PerplexityBot
Allow: /
Disallow: /wp-admin/

# Perplexity-User (real-time)
User-agent: Perplexity-User
Allow: /

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

# Googlebot - Allow
User-agent: Googlebot
Allow: /

# Bingbot
User-agent: Bingbot
Allow: /

# CCBot (Common Crawl) - Allow for open dataset
User-agent: CCBot
Allow: /
Disallow: /wp-admin/

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

# ============================================
# Bad Bots - Block
# ============================================
User-agent: Bytespider
Disallow: /

User-agent: PetalBot
Disallow: /

User-agent: AhrefsBot
Disallow: /

User-agent: SemrushBot
Disallow: /

User-agent: MJ12bot
Disallow: /

User-agent: DotBot
Disallow: /

User-agent: BLEXBot
Disallow: /

# ============================================
# Image & Mobile Crawlers
# ============================================
User-agent: Googlebot-Image
Allow: /

User-agent: Googlebot-Mobile
Allow: /

# ============================================
# Sitemaps
# ============================================
Sitemap: https://www.123bpr.online/sitemap_index.xml
Sitemap: https://www.123bpr.online/sitemap.xml
Sitemap: https://www.123bpr.online/page-sitemap.xml
Sitemap: https://www.123bpr.online/image-sitemap.xml

# HTML Sitemap for users & AI crawlers (LLM-readable)
# https://www.123bpr.online/sitemap.html

# ============================================
# IndexNow API Key (Bing, Yandex, Naver, Seznam)
# https://www.indexnow.org/
# ============================================
# IndexNow key file: https://www.123bpr.online/7c959973731b4ac6a6127a8675ed2b14.txt