serversathome/homelabhero — explained in plain English
Analysis updated 2026-05-18
Manage TrueNAS, Proxmox, and Linux servers by typing plain English requests to Claude in a browser.
Get a read-only inventory sweep of every VM, container, and app running across your home lab.
Discover new devices on your network and register the ones you pick as manageable hosts.
Keep your whole home lab command center patched and updated with a single hh update command.
| serversathome/homelabhero | ramset/pihole-ha-cluster | goatstarter/goat-fable | |
|---|---|---|---|
| Stars | 56 | 56 | 57 |
| Language | Shell | Shell | Shell |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 4/5 | 3/5 | 3/5 |
| Audience | ops devops | ops devops | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires a fresh Ubuntu or Debian based LXC or host, the installer relies on apt, systemd, and visudo.
HomelabHero turns a bare Linux server, specifically a fresh Ubuntu LXC container, into a command center for managing a home lab. Running one install command sets up Claude Code, the AI coding assistant, along with a web interface called claudecodeui, preloaded with troubleshooting knowledge, and connects everything to a credential broker. That broker lets Claude control other machines on the network, like TrueNAS storage servers, Proxmox virtualization hosts, and plain Linux boxes, over SSH, without Claude ever seeing the actual passwords or keys used to log in. After installation, the normal way to use HomelabHero is through a browser. You type plain language requests such as "how is everything doing" or "restart jellyfin" and Claude carries them out. Power users can also use a set of terminal commands, prefixed with hh, to list registered hosts, run commands on them, check connectivity, view what is running across all machines, discover new hosts on the network, and keep the whole install patched and current with a single update command. The security design is the standout feature. Three separate user accounts are involved: the person installing HomelabHero, a low privilege account that actually runs Claude, and a vault account that owns every stored credential. Claude can only reach another machine by calling a narrow broker program, which looks up the right credential in the vault and makes the connection, but never hands the credential itself back to Claude. Every command run through this broker is logged to a file Claude cannot read or erase, so even if Claude were somehow compromised, it could not steal secrets or hide what it did. This does not limit what Claude can run on a machine it is already permitted to reach, so approval prompts inside the tool are still a second layer of protection. The project only supports Ubuntu and Debian based systems, since the installer relies on apt, systemd, and visudo, and it has not been tested on other distributions. It is aimed at people running home labs who want an AI assistant managing their infrastructure without exposing credentials directly to that assistant.
HomelabHero installs Claude Code plus a web UI on a Linux server and lets it manage your other home lab machines over SSH through a credential broker that never exposes passwords to the AI.
Mainly Shell. The stack also includes Shell, Claude Code, Node.js.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
Verify against the repo before relying on details.