Ollama

Model manager and local server. Downloads, stores, and runs LLMs, and exposes a local API for front ends to talk to.

Open WebUI

Browser-based chat front end. Auto-detects models running in Ollama and gives them a familiar chat interface.

Claude Code

The engineering assistant building and documenting the lab itself — installing software, writing this site, keeping the journal.

Tailscale

Private mesh VPN connecting the lab's nodes securely, without exposing anything to the public internet.

See the Engineering Notebook for install-time observations on each of these, and Architecture for how they fit together across both nodes.