Home & network
Control the physical systems that make the homelab useful.
unifiprotectassistantA consistent, scriptable command surface across home automation, network infrastructure, media, cloud tools, and operations — built for you at the terminal and for local agents behind the scenes.
Click to copy · installs the single binary on macOS and Linux
$ home overview ops --json
{ "ok": true, "modules": 17 }
17 modules
Every adapter keeps the service-specific detail where it belongs, then exposes the same command shape, structured output, and error model at the surface.
Control the physical systems that make the homelab useful.
unifiprotectassistantSearch, play, group, and announce across the house.
spotifysonosttsRead the services that connect home life to the wider world.
googlegmailgcalgdrivediscordInspect code, stacks, planning, and deployments from one tool.
githubgraphitelinearvercelBring infrastructure health and availability into one briefing.
beszeluptime-kumaOne interface
The module contract is the source of truth for commands, configuration, status, and generated agent skills. That means a new integration plugs into the same behavior instead of inventing another mini-CLI.
human-readable output · --json · completion
commands · configuration · generated skills
authentication · API quirks · normalized errors
Operations
The useful unit is often not a single API call. home can combine service health, infrastructure pressure, alerts, and availability into the briefing you actually wanted.
$ home overview ops
NETWORK UniFi controller and sites reachable
SERVICES Home Assistant and Protect ready
SYSTEMS Beszel hosts reporting normally
AVAILABILITY Uptime Kuma has no active incidents
ready no action requiredTurn module status into one operational view instead of checking every dashboard by hand.
Every command can emit JSON, so a shell pipeline and a local model consume the same result.
Read widely, then require an intentional --yes before guarded commands can change external state.
Agent-ready
home is designed around the awkward parts of local automation: credentials, changing APIs, partial failures, and actions that should never happen silently.
The same manifests that define the CLI generate one focused skill per module, so agent instructions do not drift from the actual command surface.
Configuration lives under one root and supports the platform keyring, encrypted files, or environment-backed secrets without putting credentials in prompts.
Stable JSON, error codes, and explicit mutation guards let an agent recover from failures and know when it must ask before acting.
Get started
Install one binary, configure only the modules you use, and generate the local skills when you want an agent to operate through the same interface.