A website should be files on a disk, not a server somebody keeps alive.

What this involves

• Static output instead of a running application

  Every page is rendered at build time and served as a file. There is no application server to keep alive behind a company website, no session store, and at request time nothing for an attacker to reach.

• Two languages whose routes belong together

  Both language versions live in one route table, so every page knows its counterpart. hreflang, the language switch and the sitemap are generated from that table instead of being maintained in three places that drift apart.

• Structured data and sitemap from the same source

  schema.org markup as JSON-LD per page type and a generated sitemap, both configured to the specification you give us rather than to the default of an extension. What a machine reads and what a person reads come out of one file.

• Content as data, pages as thin wrappers

  Texts live in typed data files and the page is only the wrapper around them. Adding a page means adding data, which is why a site of this shape is still maintainable by whoever comes after the person who set it up.

• What the build produces, and how it reaches the server

  Image variants and self-hosted fonts are produced during the build, so nothing is fetched from a third party while a visitor waits. Delivery is copying the resulting directory into the web root; going back a version is putting the previous directory back.