# Synapse Global Theranostics - Robots.txt
# https://synapseglobaltheranostics.com
# Last Updated: 2025-01-21

# ===========================================
# SEARCH ENGINE CRAWLERS
# ===========================================

# Google
User-agent: Googlebot
Allow: /
Crawl-delay: 1

User-agent: Googlebot-Image
Allow: /

User-agent: Googlebot-Video
Allow: /

User-agent: Googlebot-News
Allow: /

User-agent: Storebot-Google
Allow: /

# Bing & Microsoft
User-agent: Bingbot
Allow: /
Crawl-delay: 1

User-agent: MSNBot
Allow: /

User-agent: AdIdxBot
Allow: /

# Yahoo
User-agent: Slurp
Allow: /

# DuckDuckGo
User-agent: DuckDuckBot
Allow: /

# Baidu (China)
User-agent: Baiduspider
Allow: /

# Yandex (Russia)
User-agent: YandexBot
Allow: /

User-agent: YandexImages
Allow: /

# Naver (Korea)
User-agent: Yeti
Allow: /

# Sogou (China)
User-agent: Sogou
Allow: /

# Qwant (Europe)
User-agent: Qwantify
Allow: /

# Ecosia
User-agent: Ecosia
Allow: /

# ===========================================
# SOCIAL MEDIA CRAWLERS
# ===========================================

User-agent: Twitterbot
Allow: /

User-agent: facebookexternalhit
Allow: /

User-agent: LinkedInBot
Allow: /

User-agent: Pinterest
Allow: /

User-agent: Pinterestbot
Allow: /

User-agent: Slackbot
Allow: /

User-agent: Slackbot-LinkExpanding
Allow: /

User-agent: WhatsApp
Allow: /

User-agent: TelegramBot
Allow: /

User-agent: Discordbot
Allow: /

# ===========================================
# AI & LLM CRAWLERS (AI GEO Optimization)
# ===========================================

# OpenAI / ChatGPT
User-agent: GPTBot
Allow: /
# GPTBot crawls for training and plugins

User-agent: ChatGPT-User
Allow: /
# ChatGPT browsing feature

# Anthropic / Claude
User-agent: Claude-Web
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: Anthropic
Allow: /

# Google AI / Bard / Gemini
User-agent: Google-Extended
Allow: /
# Google-Extended is for Bard/Gemini training

# Perplexity AI
User-agent: PerplexityBot
Allow: /

User-agent: Perplexity-User
Allow: /

# Cohere AI
User-agent: cohere-ai
Allow: /

User-agent: Cohere-ai
Allow: /

# You.com AI
User-agent: YouBot
Allow: /

# Neeva AI (acquired by Snowflake)
User-agent: NeevaBot
Allow: /

# Meta AI
User-agent: FacebookBot
Allow: /

User-agent: Meta-ExternalAgent
Allow: /

User-agent: Meta-ExternalFetcher
Allow: /

# Apple / Siri
User-agent: Applebot
Allow: /

User-agent: Applebot-Extended
Allow: /

# Amazon / Alexa
User-agent: Amazonbot
Allow: /

# Microsoft Copilot
User-agent: Copilot
Allow: /

# Hugging Face
User-agent: HuggingFaceBot
Allow: /

# Common Crawl (used by many AI models)
User-agent: CCBot
Allow: /

# AI2 (Allen Institute for AI)
User-agent: AI2Bot
Allow: /

# Diffbot (knowledge graph)
User-agent: Diffbot
Allow: /

# Brave Search AI
User-agent: BraveBot
Allow: /

# Mojeek (independent search)
User-agent: MojeekBot
Allow: /

# Phind (developer AI search)
User-agent: Phind
Allow: /

# Kagi Search
User-agent: Kagi
Allow: /

# ===========================================
# SEO & ANALYTICS TOOLS
# ===========================================

User-agent: AhrefsBot
Allow: /
Crawl-delay: 5

User-agent: SemrushBot
Allow: /
Crawl-delay: 5

User-agent: MJ12bot
Allow: /
Crawl-delay: 5

User-agent: DotBot
Allow: /

User-agent: rogerbot
Allow: /

User-agent: Screaming Frog SEO Spider
Allow: /

# ===========================================
# GENERAL RULES
# ===========================================

# Default rule for all other crawlers
User-agent: *
Allow: /
Crawl-delay: 2

# ===========================================
# DISALLOWED PATHS
# ===========================================

# Block development and private files
User-agent: *
Disallow: /*.json$
Disallow: /*.xml$
Disallow: /api/
Disallow: /.well-known/
Disallow: /private/
Disallow: /admin/
Disallow: /tmp/

# Allow specific files for AI discovery
User-agent: *
Allow: /llms.txt
Allow: /robots.txt
Allow: /sitemap.xml

# ===========================================
# AI-SPECIFIC CONTENT HINTS
# ===========================================

# LLMs.txt location for AI assistants
# See: https://llmstxt.org/
# File: /llms.txt

# ===========================================
# SITEMAP & HOST
# ===========================================

# Primary sitemap index (contains all sitemaps)
Sitemap: https://synapseglobaltheranostics.com/sitemap-index.xml

# Individual sitemaps (also accessible directly)
Sitemap: https://synapseglobaltheranostics.com/sitemap.xml
Sitemap: https://synapseglobaltheranostics.com/sitemap-pages.xml
Sitemap: https://synapseglobaltheranostics.com/sitemap-blog.xml
Sitemap: https://synapseglobaltheranostics.com/sitemap-images.xml
Sitemap: https://synapseglobaltheranostics.com/sitemap-ai.xml

Host: https://synapseglobaltheranostics.com