User-agent: *
Disallow: /wp-admin/
Disallow: /wp-includes/
Disallow: /wp-content/plugins/
Disallow: /wp-content/cache/
Disallow: /readme.html
Disallow: /trackback/
Disallow: /xmlrpc.php
Disallow: /wp-login.php
Disallow: /wp-register.php

# Interní / pomocné adresáře
Disallow: /wp-content/uploads/wp-import-export-lite/

# Vyhledávání
Disallow: /?s=
Disallow: /*?s=

# Duplicitní a málo hodnotné archivy
Disallow: /*/page/
Disallow: /author/

# Povolit assety potřebné pro rendering a indexaci
Allow: /wp-content/uploads/
Allow: /wp-content/themes/
Allow: /wp-content/plugins/*.css
Allow: /wp-content/plugins/*.js
Allow: /wp-includes/css/
Allow: /wp-includes/js/


# =========================================================
# AI / LLM crawlers
# =========================================================

# OpenAI training crawler
User-agent: GPTBot
Disallow: /

# ChatGPT browsing / retrieval crawler
User-agent: ChatGPT-User
Disallow: /

# OpenAI search crawler
User-agent: OAI-SearchBot
Disallow: /

# Anthropic
User-agent: ClaudeBot
Disallow: /

User-agent: Claude-SearchBot
Disallow: /

User-agent: Claude-User
Disallow: /

User-agent: anthropic-ai
Disallow: /

# Common Crawl
User-agent: CCBot
Disallow: /

# Google AI training
User-agent: Google-Extended
Disallow: /

# ByteDance / TikTok AI crawler
User-agent: Bytespider
Disallow: /

# Amazon AI crawler
User-agent: Amazonbot
Disallow: /

# Perplexity
User-agent: PerplexityBot
Disallow: /

User-agent: Perplexity-User
Disallow: /

# Meta / Facebook AI crawler
User-agent: meta-externalagent
Disallow: /

User-agent: meta-externalfetcher
Disallow: /

# Apple AI
User-agent: Applebot-Extended
Disallow: /

# Diffbot
User-agent: Diffbot
Disallow: /

# Imagesift
User-agent: ImagesiftBot
Disallow: /

# YouBot / You.com
User-agent: YouBot
Disallow: /


# =========================================================
# Sitemap
# =========================================================

Sitemap: https://tht.cz/sitemap_index.xml