Skip to content
Shll.me

Article

How Geekity organizes blog post files and site data

Explains how Geekity organizes blog post files and site data, detailing directory structures, site configuration, and theme management. Includes instructions for modifying tags and retrieving different content representations via curl requests.

This post is the file content/posts/2026-01-01-hello-world.md. The files are the source of truth; the SQLite index under data/ is derived from them, so deleting it is safe and the next boot rebuilds it.

Where things are

  • content/posts/ and content/pages/ hold the documents. Everything else in content/ is left alone.
  • content/_data/site.json is the site title, tagline, author and how many posts a listing page holds. It is the site global in every template.
  • themes/ does not exist yet. A theme is a directory in there with a theme.json in it, overriding one template at a time; anything it does not override still comes from the package. theme in site.json says which one the site wears, and no theme at all is the one the package ships.
  • geekity.config.ts is the port, the directories and the public base URL.

Two things to try

Add a tag to the front matter above and reload /tag/introductions/. Then ask this page for its other representations:

curl -H 'Accept: text/markdown' http://localhost:3000/2026/01/hello-world/
curl http://localhost:3000/2026/01/hello-world/index.json

Same URL, three bodies.