A blog with no server

The site you're reading is a folder of Markdown files. A script turns them into HTML, the HTML goes on a CDN, and that's the whole architecture. There's no database to back up, no server to patch, no admin panel getting broken into while I'm not looking.

I arrived here after years of maintaining precisely the opposite.

What I gave up

Comments. Most of a blog's complexity doesn't come from the writing, it comes from what other people write underneath it. Moderation, spam, GDPR, a database that has to exist. I gave it up. If somebody wants to respond, they write to me.

Real-time analytics. I don't need to know how many people are reading right now. A number per month is plenty, and I can have that without putting a tracker in someone else's browser.

A CMS that runs. This site's editor is static HTML: it opens at /admin/, writes to a Git repo through the GitHub API, and has no process of its own running anywhere. It can't go down, because it never comes up.

What I got

Versioning for free — every piece has a history, every change is a commit, nothing is ever lost. Zero cost. And, more important than either: nothing to patch. A static site has no attack surface in the usual sense of the phrase.

The real trade is that every dynamic feature becomes a separate decision rather than something that arrives by default. Forms, search, comments — each one makes you look at it and ask whether it's worth it. Most of the time it isn't.

When I wouldn't do this

If you need user accounts, content that changes every few minutes, or dozens of authors who will never touch Git — then a static site is a fight against the tool. For a personal blog, it's exactly the right one.