2026.09.04

Installs both local-model runtimes. Ollama's server runs as a background service, so port 11434 answers in a fresh terminal and after a reboot. LM Studio's lms command is put on your PATH, which the application does not do itself.

Neither downloads a model. Both stores start empty; ollama pull and lms get are yours to run. A single model is often larger than everything else this installs put together.

Downloads about 4.71 GB, measured 2026-09-10 on macOS 26.5.2, Apple Silicon.