How this site works

I have a busy day job, so I can’t hand-write and hand-test every guide on this site. Instead, I use AI to help, and I’d rather tell you exactly how than pretend otherwise.

How a guide gets made

Research. Each guide starts from the official documentation, release notes and changelogs for the tool it covers. Sources are linked at the bottom of every post.
Draft. AI (Claude, made by Anthropic) writes the first draft from that research.
Run the commands. Where a guide uses Linux commands, they’re run automatically on a fresh test machine. If they fail, the guide doesn’t go out until it’s fixed.
I review it. I read every post before it’s published. Nothing goes live without that.

What the labels mean

Every guide has a small card at the top that tells you how far it’s been checked.

Commands run on [system]: the commands were executed on a test machine on the date shown, and worked.
From my own setup: a story or walkthrough of something I actually did on my own kit, like moving this site off WordPress.
Based on the official docs: built from the vendor’s documentation but not run on the real hardware (for example pfSense or Synology). Take a backup and follow along carefully.
Published / Last updated: an older guide from before this process. It hasn’t been re-checked yet, so commands or versions may have moved on. These are being worked through, most-read first.

What I won’t do

Claim something was tested when it wasn’t.
Publish anything I haven’t read.
Hide sponsored or affiliate links. They’re always labelled.

Found a mistake?

AI and I both get things wrong sometimes. If a guide doesn’t work for you, tell me. I’ll fix it and add a short corrections note at the bottom of the post, so you can see what changed.