# Allow all robots complete access
User-agent: *
Allow: /

Disallow: /api/
Disallow: /cp/
Disallow: /!/

# Legacy language-switch parameter URLs: redirect-only duplicates of the clean
# localized URLs (the switcher now links directly). Blocked to save crawl budget.
Disallow: /*set_locale=

# Sitemaps
Sitemap: https://vodlix.com/sitemap.xml
Sitemap: https://vodlix.com/fr/sitemap.xml
Sitemap: https://vodlix.com/es/sitemap.xml
Sitemap: https://vodlix.com/ar/sitemap.xml
Sitemap: https://vodlix.com/nl/sitemap.xml
Sitemap: https://vodlix.com/zh/sitemap.xml
Sitemap: https://vodlix.com/de/sitemap.xml
Sitemap: https://vodlix.com/pt/sitemap.xml
Sitemap: https://vodlix.com/ja/sitemap.xml
Sitemap: https://vodlix.com/it/sitemap.xml
Sitemap: https://vodlix.com/tr/sitemap.xml
Sitemap: https://vodlix.com/pl/sitemap.xml
Sitemap: https://vodlix.com/video-sitemap.xml

# AI discovery: MCP metadata is served at /.well-known/mcp.json (standard URL; not a robots.txt field)
#
# EntityMap v1.0 (https://entitymap.org/spec/v1.0) — a structured, attributed
# index of what this site knows, for AI systems. Where the sitemaps above list
# pages, these list entities and the evidence behind them. Root URLs per locale
# (standard paths; EntityMap defines no robots.txt field either):
#   https://vodlix.com/entitymap.json      + /entitymap.html
#   https://vodlix.com/<locale>/entitymap.json  for fr, es, ar, nl, zh, de, pt, ja, it, tr, pl
#
# Other structured endpoints, so nothing needs to parse our HTML:
#   https://vodlix.com/llms.txt        navigation and citation guide
#   https://vodlix.com/llms-full.txt   every published page as Markdown, one file (per locale too)
#   https://vodlix.com/faqs.json       deduplicated question/answer pairs (per locale too)
#   <any page URL>.md                  that page as Markdown (e.g. /blog/ai-in-streaming.md)