System tools

Installers, Storage, Migration Hosts, and Routing keep the host healthy and connected. Check these when downloads fail, disk fills up, or you need to move a workload.

Installers (Setup)

Installers (Setup) verifies local binaries used for catalog search and inference.

Each card should show a green check and Ready. Paths live under the app data directory (or Homebrew for MLX). Use Refresh after a manual install. If a card is not ready, install from the card actions before browsing Advanced models.

Tool	Purpose
llmfit	Catalog search and host-fit scoring
llama.cpp (llama-server)	Local inference server for loaded models
MLX (mlx_lm)	Apple Silicon MLX server when available

Storage

Storage shows what LaunchPad has cached:

• Images — OS images and size on disk
• Models — model artifacts with status In use or Available

Delete with the trash icon per row. Prune expired clears stale caches; Delete all is destructive — unload running models first if status is In use.

Migration Hosts

Targets for Migrate (from an instance or composition menu):

• Hosts on the same network appear automatically (On network).
• Add host for a machine on another network or one that discovery missed.
• Refresh re-scans.

Use this when you want to move a VM or composition to another LaunchPad host without recreating from scratch.

Routing

Routing lists host TCP ports forwarded to guest instance IPs. Bind is localhost by default.

1. Click Add port forward.
2. Choose host port, guest IP / port, and bind address.
3. Refresh after changes outside the UI.

Use forwards when a service inside a VM should be reachable as localhost:PORT on the Mac (or another bind you choose).

Help

Help → View documentation opens the published docs (this site). Prefer the in-app link when you are already troubleshooting in the GUI.

Next

• Settings — startup, Multipass visibility, appearance
• Getting started — back to Home