# ======================================
# Robots.txt for Trulieve.com
# Last updated: Oct 23, 2025
# Version: 03
# ======================================

Sitemap: https://www.trulieve.com/sitemap-index.xml

# ======================================
# Global Rules
# ======================================

User-agent: *

# --- Block low-value pages/paths
Disallow: /*search*
Disallow: /*?filter=*
Disallow: /*?url_path=*
Disallow: /*?page=*
Disallow: /*account*
Disallow: /*checkout*
Disallow: /*login*
Disallow: /*settings*
Disallow: /*patient*
Disallow: /*wishlist*
Disallow: /*profile*
Disallow: /*thank-you*

# --- Allow important pages/paths
Allow: /*graphql
Allow: /*static

# ======================================
# Google Special Bots
# ======================================
User-agent: Mediapartners-Google
Allow: /*media

User-agent: AdsBot-Google
Allow: /*media

# ======================================
# Restricted Bots (blocked)
# ======================================
User-agent: Baiduspider
Disallow: /

User-agent: 360Spider
Disallow: /

User-agent: Yisouspider
Disallow: /

User-agent: PetalBot
Disallow: /

User-agent: Bytespider
Disallow: /

User-agent: Sogou web spider
Disallow: /

User-agent: Sogou inst spider
Disallow: /

User-agent: Robozilla
Disallow: /

User-agent: yahoo-mmcrawler
Disallow: /

User-agent: psbot
Disallow: /

User-agent: asterias
Disallow: /

User-agent: yahoo-blogs/v3.9
Disallow: /

User-agent: Yandex
Disallow: /

User-agent: BLEXBot
Disallow: /

# ======================================
# Crawl Delays (non-Google bots)
# ======================================
User-agent: rogerbot
Crawl-delay: 10

User-agent: dotbot
Crawl-delay: 10

User-agent: AhrefsBot
Crawl-delay: 10

User-agent: MJ12bot
Crawl-delay: 10

User-agent: Pinterestbot
Crawl-delay: 1

# ======================================
# Sitemap (repeat for redundancy)
# ======================================
Sitemap: https://www.trulieve.com/sitemap-index.xml