Ollama

#Introduction

Ollama is a client for artificial intelligence models, so it is the foundation upon which you then install an AI that you want to use.

Ollama has two distinctive features. The first is that it lets you use an AI locally. This means that instead of going to a company’s AI chat page, the model is installed on your computer and you use it directly without visiting any website.

This benefits you in three ways. First, because the data of everything you do stays on your PC, so no company uses it. Second, because you can use the AI without an internet connection, and third, because you can bypass any censorship that an AI model you are using on a website might have.

And the second distinctive feature is that it works through your computer’s terminal, or the command prompt on Windows. This means you don’t have to use a separate application. When you install Ollama, you’ll then have to use your device’s console to install and run the model you want, and you write the questions and prompts in the console, where you’ll also get the answers.

#Get Started

Create a vGPU Linux desktop on Isard.

Paste this in the terminal to install Ollama:

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

Run ollama in your terminal to open the interactive menu:

ollama

Navigate with ↑/↓, press enter to launch, → to change model, and esc to quit.

The menu provides quick access to:

• Run a model – Start an interactive chat
• Launch tools – Claude Code, Codex, OpenClaw, and more
• Additional integrations – Available under “More…”

#Commands

List the models you have installed:

ollama list

If the Ollama server is running on port 11434:

ollama serve