Self-Hosting

The Always-On AI: What Happens When You Leave Claude Code Running

There’s a moment that changes how you think about AI assistants. It’s not the first clever answer, or the time it writes a function you were dreading. It’s quieter than that.

It’s the moment you realize you haven’t opened a terminal in three days — and your server is running better than ever.

That’s what happens when you stop treating Claude Code as a tool you pick up and put down, and start letting it run. Persistently. Always on. Waiting.

Controlling My Home Server From Telegram With Claude Code

I run a home server called rocklab — Ubuntu 24.04, a pile of Docker containers, and Claude Code acting as my on-call IT department. It handles routine maintenance, helps me publish blog posts, and executes whatever tasks I throw at it.

The one missing piece was mobility. If I wanted to check on something or kick off a task, I had to be at a terminal. That changed today when I set up Claude Code Channels — specifically the Telegram plugin — which lets me message my server from anywhere and have Claude respond like a proper remote assistant.