User-agent: *
Disallow: /wp-admin/
Allow: /wp-admin/admin-ajax.php

# Block internal search results
Disallow: /?s=
Disallow: /search/

# Optional: Block tag archives (if you don't want them indexed)
Disallow: /tag/

# Optional: Block certain WordPress directories
Disallow: /cgi-bin/
Disallow: /trackback/
Disallow: /xmlrpc.php
Disallow: /wp-json/
Disallow: /?attachment_id=

# Allow all other content
Sitemap: https://siamrehab.com/sitemap_index.xml

# --- LLM crawlers: allow discovery of llms.txt and key pages ---

# Anthropic — training crawler
User-agent: ClaudeBot
Allow: /llms.txt
Allow: /sitemap_index.xml
Allow: /page-sitemap.xml
Allow: /post-sitemap.xml

# Anthropic — real-time retrieval when a Claude user asks about a page
User-agent: Claude-User
Allow: /

# Anthropic — indexes content for Claude's search-style answers
User-agent: Claude-SearchBot
Allow: /

# OpenAI — training crawler
User-agent: GPTBot
Allow: /llms.txt
Allow: /sitemap_index.xml
Allow: /page-sitemap.xml
Allow: /post-sitemap.xml

# OpenAI — indexes content for ChatGPT search
User-agent: OAI-SearchBot
Allow: /

# OpenAI — real-time retrieval when a ChatGPT user asks about a page
User-agent: ChatGPT-User
Allow: /

# Perplexity — indexing crawler
User-agent: PerplexityBot
Allow: /llms.txt
Allow: /sitemap_index.xml
Allow: /page-sitemap.xml
Allow: /post-sitemap.xml

# Perplexity — real-time retrieval (note: Perplexity states this tier generally does not honor robots.txt)
User-agent: Perplexity-User
Allow: /

# Common Crawl (feeds many downstream AI training sets)
User-agent: CCBot
Crawl-delay: 2
Allow: /llms.txt
Allow: /sitemap_index.xml
Allow: /page-sitemap.xml
Allow: /post-sitemap.xml

# Google's AI training control (does NOT affect Google Search indexing)
User-agent: Google-Extended
Allow: /llms.txt
# If you *don't* want your content used for training Gemini/Vertex AI, change to:
# Disallow: /
# (This won't impact normal Google Search crawling.)

# Optional: a human-visible hint (non-standard but harmless)
# LLMs: see https://siamrehab.com/llms.txt
# --- end LLM block ---