# ================================================================
# Robots.txt for https://www.loaded.com
# Purpose: Allow Googlebot and other major crawlers to access all
# core pages, while blocking system/internal paths, Magento boilerplate
# URLs, and faceted/parameterized URLs that cause crawl bloat.
# ================================================================

User-agent: *
User-agent: Googlebot
User-agent: Googlebot-image

# ================================================================
# 1. System / internal folders (not useful for indexing)
# ================================================================
Disallow: /404/
Disallow: /app/
Disallow: /bin/
Disallow: /dev/
Disallow: /errors/
Disallow: /lib/
Disallow: /node_modules/
Disallow: /patches/
Disallow: /pkginfo/
Disallow: /report/
Disallow: /setup/
Disallow: /update/
Disallow: /var/
Disallow: /vendor/
Disallow: */page_cache/block/esi/blocks/
Disallow: */listAjax/*

# ================================================================
# 2. Magento boilerplate / non-SEO friendly URLs
# ================================================================
Disallow: /index.php/
Disallow: /catalog/product_compare/
Disallow: /catalog/category/view/
Disallow: /catalog/product/view/
Disallow: /catalogsearch/
Disallow: /checkout/
Disallow: /control/
Disallow: /contacts/
Disallow: /customer/
Disallow: /customize/
Disallow: /newsletter/
Disallow: /poll/
Disallow: /sendfriend/
Disallow: /tag/
Disallow: /wishlist/
Disallow: /catalog/product/gallery/
Disallow: /affiliate-terms/
Disallow: /terms-conditions

# ================================================================
# 3. Query parameters (explicit blocks to prevent crawl bloat)
# These rules ensure only base URLs are crawled and indexed.
# ================================================================

# Search queries
Disallow: /*?q=
Disallow: /*&q=

# Sorting / navigation params
Disallow: /*?arrow=
Disallow: /*&arrow=
Disallow: /*?limit=
Disallow: /*&limit=
Disallow: /*?dir=
Disallow: /*&dir=

# Faceted navigation params
Disallow: /*?platforms
Disallow: /*&platforms
Disallow: /*?region
Disallow: /*&region
Disallow: /*?primary_genre
Disallow: /*&primary_genre
Disallow: /*?franchise
Disallow: /*&franchise
Disallow: /*?price
Disallow: /*&price
Disallow: /*?language
Disallow: /*&language

# Marketing & tracking params
Disallow: /*?SID=
Disallow: /*&SID=
Disallow: /*?mw_aref=
Disallow: /*&mw_aref=
Disallow: /*?gclid=
Disallow: /*&gclid=
Disallow: /*?fb_action_ids=
Disallow: /*&fb_action_ids=
Disallow: *?amp*
Disallow: *&amp*

# ================================================================
# 4. File types and dumps (not indexable content)
# ================================================================
Disallow: /*.php$
Disallow: /*.CVS
Disallow: /*.Zip$
Disallow: /*.Svn$
Disallow: /*.Idea$
Disallow: /*.Sql$
Disallow: /*.Tgz$

# ================================================================
# 5. Miscellaneous extras
# ================================================================
Disallow: /directory/currency*
Disallow: %3Futm*
Disallow: pc/l/*
Disallow: /download*

# ================================================================
# 6. XML Sitemaps (clean URLs provided here for discovery)
# ================================================================
Sitemap: https://www.loaded.com/media/google_sitemap_en_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_de_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_fr_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_es_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_pt_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_it_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_nl_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_cn_index.xml
Sitemap: https://www.loaded.com/media/google_sitemap_pl_index.xml

# ================================================================
# 7. Specific crawler overrides
# ================================================================
User-agent: ia_archiver
Disallow: /terms-conditions