Kodex: reclaiming a space for writing

Published on October 25, 2026 3 min read 420 words

Kodex is the project behind Kazztor.net, a personal publication for sharing knowledge, ideas, and experiences. Its purpose is to give ideas room to develop, with a presentation that pays attention to readability, typography, and reading rhythm. Depth matters; length depends on what each piece needs.

A project put on hold

In its early stages, the site was a place to explore the reading experience. Its articles contained sample text used to assess font sizes, spacing, and composition. They were not an editorial collection ready for publication.

After several months without maintenance, the repository held decisions from different stages: older deployment settings, dependencies that needed review, and features that were not fully connected. Resuming the project meant understanding what was already there before adding more.

Recovering a clear direction

Recovery began with a read-only audit. Git, architecture, dependencies, themes, languages, images, search, and navigation were examined. That review separated verified behavior from intentions that existed only in documentation.

Using Repositorio DEV 0.3 as a guide, the next task was to define the product. Kazztor.net was confirmed as the domain, Spanish as the base language, and English as the second language. The project would remain a static publication, with content stored alongside code and deployment targeting Hetzner through Nginx. Features such as a newsletter were left outside the current scope.

Building a consistent foundation

Technical recovery moved forward in increments: an environment using Node.js 24 and npm 11, Astro 7, and a bilingual editorial model. Each translation has its own route and is connected to its counterpart through an explicit key. Markdown is the default format; MDX remains available for compositions that require components.

The presentation adopted style-foundation v0.2.0 and light, dark, and system theme preferences. Editorial navigation, categories, RSS, and metadata were aligned with the active language. Pagefind connected search to published articles, while automated tests began covering both representative content and empty collections.

The latest work focused on responsive local images, alternative text, reserved image dimensions, and tables and code blocks that remain usable on small screens. Checks also confirmed that essential content remains accessible without JavaScript.

Returning to writing

The recovery aims to reduce the effort required to publish and maintain the site. Reviewing the legal text and verifying deployment and rollback on the actual infrastructure still remain.

This article marks a return to original content. The technical foundation exists to support that work: writing, reviewing, translating, and sharing pieces that offer something useful to their readers.