# Crawlers Setup
User-agent: *

# Website Sitemap
Sitemap: https://www.sugarfreeshops.com/media/sitemap/google_sitemap_en.xml
Sitemap: https://www.sugarfreeshops.com/media/sitemap/google_sitemap_el.xml
#


# Allowable Index
Allow: /*?p=
Allow: /media/

# Directories
Disallow: /app/
Disallow: /bin/
Disallow: /dev/
Disallow: /lib/
Disallow: /var/
Disallow: /404/
Disallow: /js/
#Disallow: /media/
Disallow: /phpserver/
Disallow: /pkginfo/
Disallow: /report/
Disallow: /setup/
Disallow: /update/
Disallow: /vendor/
Disallow: /cgi-bin/
Disallow: /downloader/
Disallow: /errors/
Disallow: /includes/
Disallow: /scripts/
Disallow: /shell/
Disallow: /skin/
Disallow: /stats/

# Paths (clean URLs)
Disallow: */index.php/
Disallow: */catalog/product_compare/
Disallow: */catalog/category/view/
Disallow: */catalog/product/view/
Disallow: */catalog/product/gallery/
Disallow: */catalogsearch/
Disallow: */search/
#Disallow: */checkout/
Disallow: */control/
Disallow: */contacts/
Disallow: */customer/
Disallow: */customize/
Disallow: */newsletter/
Disallow: */review/
Disallow: */tag/
Disallow: */sendfriend/
Disallow: */wishlist/
Disallow: *//*

# Files
Disallow: /composer.json
Disallow: /composer.lock
Disallow: /CONTRIBUTING.md
Disallow: /CONTRIBUTOR_LICENSE_AGREEMENT.html
Disallow: /COPYING.txt
Disallow: /Gruntfile.js
Disallow: /LICENSE.txt
Disallow: /LICENSE_AFL.txt
Disallow: /nginx.conf.sample
Disallow: /package.json
Disallow: /php.ini.sample
Disallow: /RELEASE_NOTES.txt

# Do not index pages that are sorted or filtered.
Disallow: /*?*product_list_mode=
Disallow: /*?*product_list_order=
Disallow: /*?*product_list_limit=
Disallow: /*?*product_list_dir=

# Paths (no clean URLs)
#Disallow: /*.js$
#Disallow: /*.css$
#Disallow: /*?
#Disallow: /*?SID=
Disallow: /*.php$
Disallow: /*?p=*&

# CVS, SVN directory and dump files
Disallow: /*.CVS
Disallow: /*.Zip$
Disallow: /*.zip$
Disallow: /*.Svn$
Disallow: /*.svn$
Disallow: /*.Idea$
Disallow: /*.idea$
Disallow: /*.Sql$
Disallow: /*.sql$
Disallow: /*.Tgz$
Disallow: /*.tgz$


# Google Image Crawler Setup
User-agent: Googlebot-Image
Disallow:

User-agent: AhrefsBot
Disallow: /
Disallow: /lib/
Disallow: /*.php$
Disallow: /pkginfo/
Disallow: /report/
Disallow: /var/
Disallow: /catalog/
Disallow: /customer/
Disallow: /sendfriend/
Disallow: /review/
Disallow: /*SID=
Disallow: /*?

# Disable checkout & customer account
Disallow: /checkout/
Disallow: /onestepcheckout/
Disallow: /customer/
Disallow: /customer/account/
Disallow: /customer/account/login/

# Disable Search pages
Disallow: /catalogsearch/
Disallow: */search/
Disallow: /catalog/product_compare/
Disallow: /catalog/category/view/
Disallow: /catalog/product/view/

# Disable common folders
Disallow: /app/
Disallow: /bin/
Disallow: /dev/
Disallow: /lib/
Disallow: /phpserver/
Disallow: /pub/

# Disable Tag & Review (Avoid duplicate content)

Disallow: /tag/
Disallow: /review/

# Common files
Disallow: /composer.json
Disallow: /composer.lock
Disallow: /CONTRIBUTING.md
Disallow: /CONTRIBUTOR_LICENSE_AGREEMENT.html
Disallow: /COPYING.txt
Disallow: /Gruntfile.js
Disallow: /LICENSE.txt
Disallow: /LICENSE_AFL.txt
Disallow: /nginx.conf.sample
Disallow: /package.json
Disallow: /php.ini.sample
Disallow: /RELEASE_NOTES.txt

# Disable sorting (Avoid duplicate content)
Disallow: /*?*product_list_mode=
Disallow: /*?*product_list_order=
Disallow: /*?*product_list_limit=
Disallow: /*?*product_list_dir=

# Disable version control folders and others
Disallow: /*.git
Disallow: /*.CVS
Disallow: /*.Zip$
Disallow: /*.Svn$
Disallow: /*.Idea$
Disallow: /*.Sql$
Disallow: /*.Tgz$


#sem https://www.semrush.com/bot/

#To block SEMrushBot from crawling your site for different SEO and technical issues:
User-agent: SemrushBot-SA
Disallow: /
#To block SEMrushBot from crawling your site for Backlink Audit tool:
User-agent: SemrushBot-BA
Disallow: /
#To block SEMrushBot from crawling your site for On Page SEO Checker tool and similar tools:
User-agent: SemrushBot-SI
Disallow: /
#To block SEMrushBot from checking URLs on your site for SWA tool:
User-agent: SemrushBot-SWA
Disallow: /
#To block SEMrushBot from crawling your site for Content Analyzer and Post Tracking tools:
User-agent: SemrushBot-CT
Disallow: /
#To block SEMrushBot from crawling your site for Brand Monitoring:
User-agent: SemrushBot-BM
Disallow: /
#To block SEMrushBot from crawling your site for SEO A/B Testing tool:
User-agent: SemrushBot-SEOAB
Disallow: /

# === 1. Default Allow ===
User-agent: *
Allow: /

# === 2. Google Crawlers (Search + Gemini AI) ===
User-agent: Googlebot
Allow: /
User-agent: Googlebot-Image
Allow: /
User-agent: GoogleOther
Allow: /
User-agent: Google-Extended
Allow: /

# === 3. OpenAI / ChatGPT ===
User-agent: GPTBot
Allow: /
User-agent: ChatGPT-User
Allow: /

# === 4. Meta AI / Facebook / Instagram ===
User-agent: FacebookBot
Allow: /
User-agent: Facebot
Allow: /
User-agent: Meta-ExternalBot
Allow: /
User-agent: MetaBot
Allow: /

# === 5. Anthropic Claude ===
User-agent: anthropic-ai
Allow: /
User-agent: ClaudeBot
Allow: /

# === 6. Perplexity AI ===
User-agent: PerplexityBot
Allow: /

# === 7. Bing / Microsoft AI ===
User-agent: Bingbot
Allow: /
User-agent: BingPreview
Allow: /

# === 8. Other AI / Emerging Crawlers ===
User-agent: YouBot
Allow: /
User-agent: NeevaBot
Allow: /
User-agent: PhindBot
Allow: /
User-agent: DuckDuckBot
Allow: /
User-agent: KagiBot
Allow: /
User-agent: AndiBot
Allow: /
User-agent: TurnitinBot
Allow: /
User-agent: CCBot
Allow: /

#Tools Bots
User-agent: AhrefsBot
Allow: /

User-agent: Applebot
Allow: /