marius-hoegger.ch
My personal website, maintained with AI agents



Starting point
Talks, webinars, podcast episodes and blog posts were scattered across YouTube, Spotify, LinkedIn and event pages. I wanted one place that brings it all together and shows how I think about AI and what I build.
The website
Technically the site is deliberately simple: Nuxt with Nuxt Content, all content as Markdown files, built statically and served from GitHub Pages. Every merge to the main branch is published automatically. The visual motif is the hexagon, from the image grid on the home page to the timeline of all posts. Since May 2026 the site is also available in English.
AI in the workflow
More interesting than the site itself is how it is maintained. I work with Claude Code directly in the repository, and the repository is set up for it:
- Project knowledge for the agent: a
CLAUDE.mddescribes the site's structure, conventions and rules, so every session starts from the same baseline instead of exploring the code first. - Translation as a subagent: German is the source. A dedicated translation agent mirrors every file into the English folder, carries frontmatter, image paths and embedded components over unchanged and keeps the interface's language files in sync. Tone and rules live in the agent definition, not in my head.
- Skills for recurring copy: for LinkedIn there are skills that write the profile summary and the experience entries straight from the website's content. The website is the single source everything else is built from.
- Readable for AI: the site serves an
llms.txtgenerated at build time from the same content as the pages themselves. New posts appear there automatically, without anyone touching the file. - A feed that fills itself: talks, podcast episodes, blog posts and project updates flow from their own collections into one shared timeline. A new project with a launch date creates its own entry under "Latest".
What I learned
An AI works in a repository only as well as the repository lets it. Since the
rules live in CLAUDE.md, agents and skills instead of being re-explained every
session, the results are more consistent and I review more than I write. It is
the same lesson I pass on in client projects.
