Projects

Infrastructure

Home lab server

A self-hosted Linux server for media and a private AI assistant

· 2 min read

period
2026
role
Solo build
status
Running

An Ubuntu server at home that runs a Jellyfin media library and a private AI assistant I can message on Telegram. I manage it with scripts from Windows over SSH, with secrets kept off the repository.

My home lab is where I practise the infrastructure side of my job on my own hardware. It is an Ubuntu server that hosts a Jellyfin media library and a self-hosted AI assistant.

I manage it from Windows with a small Python tool built on Paramiko that connects over SSH (with optional key files), deploys scripts and runs checks remotely. Bash scripts handle media housekeeping, such as an interactive clean-up tool that shows each item’s size and asks for explicit confirmation before deleting anything.

The AI assistant runs as a systemd user service behind a model-routing gateway and talks to me through a private Telegram bot that only answers my own account. Its configuration and API keys live only on the server, in a file readable only by its owner, and are never committed. The repository keeps sanitised copies of scripts for review.

Highlights

  • Remote SSH management from Windows with Python
  • Self-hosted AI assistant as a systemd service, locked to one Telegram account
  • Secrets kept only on the server with owner-only permissions
  • Safe-by-default scripts with confirmation before destructive actions

Built with

  • Ubuntu Server
  • Jellyfin
  • Bash
  • Python (Paramiko)
  • systemd
  • OpenClaw

What it taught me

Running your own infrastructure makes security practical: least privilege, secrets handling and safe defaults stop being theory when it’s your server.

The code for this project is private, so I’ve described how it works rather than linking to it.