Local model setup
Make your Lucebox ready.
Check the hardware, install the qualified model, and start your private API. Everything runs on this machine.
MachineChecking…
R9700—
PCIe link—
ROCm—
Model profileChecking…
Disk free—
Private accessChecking…
Lucebox CLI / live output
State / ready
░██ ░██
░██ ░██ ░██ ░███████ ░███████ ░████████ ░███████ ░██ ░██
░██ ░██ ░██ ░██ ░██ ░██ ░██ ░██ ░██ ░██ ░██ ░██ ░██
░██ ░██ ░██ ░██ ░█████████ ░██ ░██ ░██ ░██ ░█████
░██ ░██ ░███ ░██ ░██ ░██ ░███ ░██ ░██ ░██ ░██ ░██
░██████████ ░█████░██ ░███████ ░███████ ░██░█████ ░███████ ░██ ░██
LUCEBOX
› $ lucebox manage --local
› Ready. Choose Check machine to begin.
▋
Apps
Use Lucebox where you already work.
Choose an app here. Lucebox starts the model and prepares its private connection for you.
01
Choose an app
Server stopped
Choose any supported app. Lucebox Connect installs it when needed, configures the private model, and opens it on this laptop.
One-time laptop setup
Enable one-click apps
A small Lucebox connector configures and opens supported apps on this laptop. It runs only when you click.
Open WebUI
One click
A familiar private chat interface in your browser.
OpenCode
One click
An open-source coding agent for your terminal.
Claude Code
One click
Use the local model from Anthropic's coding agent.
Codex
One click
Connect OpenAI's coding agent to your Lucebox.
Hermes Agent
One click
A general agent with terminal tools and local workflows.
OpenClaw
One click
A personal agent for local tools, automations, and channels.
Pi
Pi
One click
A compact terminal coding agent with local tools.
Connect Tailscale first. Any supported app can then open automatically on this laptop.