---
title: "install ollama"
---
# 01 — Install Ollama

## Download Ollama

Go to [ollama.com](https://ollama.com) and click **Download**. You'll get a `.zip` file. Double-click it, then drag the Ollama application to your Applications folder.

Open Ollama from Applications. You should see a little llama icon appear in your menu bar. That means it's running.

Verify it works by opening Terminal and running:

```bash
ollama --version
```

You should see something like `ollama version is 0.x.x`.

## The Ollama CLI

Ollama has a few important commands you'll use throughout this course:

```bash
ollama pull <model>     # Download a model
ollama list             # Show models you've downloaded
ollama run <model>      # Start an interactive chat with a model
ollama rm <model>       # Delete a model to free space
```

## Choose and Download a Model

Which model should you pick? Here's a guide:

| Your RAM | Recommended Model | Pull Command                  |
| -------- | ----------------- | ----------------------------- |
| 8GB      | `nemotron-3-nano` | `ollama pull nemotron-3-nano` |
| 16GB     | `gemma4:26b`      | `ollama pull gemma4:26b`      |
| 32GB+    | `qwen3:30b-a3b`   | `ollama pull qwen3:30b-a3b`   |

For this course, we'll use `gemma4:26b` in all examples. If you're using a different model, just substitute the name.

Pull the model:

```bash
ollama pull gemma4:26b
```

This will take a while — the model is about 15GB. Go get a coffee. ☕

When it finishes, verify:

```bash
ollama list
```

You should see `gemma4:26b` in the list.

## Memory Pressure — The Metric That Matters

Open **Activity Monitor** (find it in Spotlight with ⌘+Space). Click the **Memory** tab.

At the bottom is a graph labeled **Memory Pressure**. This tells you how hard your computer is working to manage RAM.

- **Green** = plenty of headroom. Your model fits comfortably.
- **Yellow** = getting tight. The model is using most of your RAM.
- **Red** = you're out of RAM. Everything is slow. The model is too big.

> **The goal:** pick the smartest model that keeps your memory pressure green during a chat.

Keep Activity Monitor open while you use Ollama. Watch the pressure graph as you run prompts. If it spikes to yellow or red, try a smaller model.

## What You've Learned

- Ollama runs as a background service (menu bar icon)
- You pull models with `ollama pull <name>`
- Model size matters — bigger models are smarter but need more RAM
- Memory pressure is your dashboard for picking the right model

**Next: [02 — Your First Prompt →](/run-ai-locally/01-getting-started/02-first-prompt/)**