Wraps the GroupDocs.Metadata .NET library as an MCP server so AI agents can read and strip metadata from documents. Exposes two tools: ReadMetadata pulls all properties (author, title, dates, custom fields) as JSON, and RemoveMetadata writes a cleaned copy back to disk. Runs via dnx (no install), as a global .NET tool, or in Docker. Configure storage paths through environment variables and optionally point to a license file to exit evaluation mode. Most useful when you're building document workflows that need metadata inspection or sanitization and want Claude or Copilot to handle it directly without writing extraction code.