User-agent: *
Allow: /

# App surfaces carry noindex in meta and X-Robots-Tag. They are NOT
# disallowed here on purpose: a disallowed URL is never fetched, so the
# noindex is never seen, and /studio is linked from the site-wide header.
# Blocking it is how a page ends up indexed as a bare URL with no snippet.

# APIs, OAuth, and MCP runtimes (not HTML documents)
Disallow: /api/
Disallow: /mcp/
Disallow: /mcp-studio
Disallow: /mcp-catalog

# AI search crawlers: allowed (no extra Disallow). Do not block
# GPTBot / ClaudeBot / PerplexityBot / Google-Extended here — avva
# is built for agents and wants accurate citations of public pages.

Sitemap: https://avva.chat/sitemap.xml