Install Ollama

Getting started with local LLMs (SWIB26)

Install Ollama

Ollama is the program that runs the models, and we use it in both halves of the workshop. Install it, open it once, and check the version.

Download

Get Ollama.dmg from https://ollama.com/download. You need macOS 14 or newer.

Install

Open the downloaded disk image and drag Ollama into the Applications folder.

Open

Open Ollama from Applications. If macOS blocks it, go to System Settings, then Privacy & Security, and choose Open Anyway.

Run the install script

curl -fsSL https://ollama.com/install.sh | sh

Leave the service alone

The script sets up Ollama as a background service (a systemd service) and starts it for you. Do not run ollama serve afterwards, because the service is already running.

Pick a chat window

There is no desktop app on Linux. Use the terminal, or install LM Studio for a chat window.

First run

When you open the app for the first time on macOS or Windows, Ollama shows “Welcome to Ollama” and then a “Create an account” screen.

Choose No thanks, I’ll use Ollama locally. The account is only for Ollama’s paid cloud models; you do not need it for this workshop. The next screen, “Run Ollama”, shows a terminal command. You can close that screen.

If you skipped the account screen by accident, nothing is lost. Local use never needs an account.

Verify

The app window opens with “Select a model”. In a terminal, run:

ollama --version

The command must print 0.40 or newer. We tested the workshop materials on 0.40.4.

Where models live

System	Folder
macOS and Linux (user install)	`~/.ollama/models`
Windows	`%USERPROFILE%\.ollama\models`
Linux (system service)	`/usr/share/ollama/.ollama/models`

You can change the location in Settings, under Model location.

Choose a model size for your hardware Install LM Studio