---
title: Working with Ollama
description: Learn how to connect your Ollama to Octarine to use Writing Assistant and Ask Octarine
---


Ollama is a simple way to run large language models on your own machine. This guide shows you how to install it, download a model, and connect it to Octarine so you can use AI features completely offline and privately.

## Prerequisites

Ollama works on macOS, Windows, and Linux. Just make sure your system has enough CPU, RAM, and storage to handle the models you want to run.

## Step 1: Install Ollama

Visit the [Ollama official website](https://ollama.com/) and download the installer for your operating system.

On macOS, open the `.dmg` file and follow the instructions. On Windows, run the `.exe` and complete the wizard. If you're on Linux, check the Ollama website for instructions specific to your distribution.

## Step 2: Pull a Model

After installing Ollama, you need to download a model. Open Terminal (on macOS/Linux) or Command Prompt or PowerShell (on Windows), and run this command:

```bash
ollama pull llama2
```

Replace `llama2` with whatever model you want to use. You can find a full list of available models in the [Ollama documentation](https://ollama.com/library).

## Step 3: Start the Ollama Service

In the same terminal or command prompt, start the Ollama server:

```bash
ollama serve
```

This runs the server on `http://localhost:11434` by default. Keep this terminal window open—you need the service running for Octarine to connect.

## Step 4: Connect Ollama to Octarine

Now let's hook it up to Octarine.

<Image src="https://pub-d9b2979edab5442388c14f8014e177b7.r2.dev/docs/working_with_ollama.png" />

Go to **Settings → AI → API Providers** and choose **Ollama**. Enter the server URL from the previous step (usually `http://localhost:11434`) and save the connection.

## Step 5: Start Using Your Models

<Image src="https://pub-d9b2979edab5442388c14f8014e177b7.r2.dev/docs/working_with_ollama_models.png" width={300} />

Open the Writing Assistant or Ask Octarine, click the model selector, and look for your Ollama models. Select one and you're ready to go!

## A Few Things to Know

The Ollama API only listens on `localhost` by default, so it's not exposed to the internet—everything stays on your machine.

If you want to see all the models you've downloaded, run `ollama list` in your terminal. This shows every model currently available.

When you're done, you can stop the Ollama service by going back to the terminal where `ollama serve` is running and pressing `Ctrl+C`.

#### Quick Answers

### Can Octarine use Ollama?

Yes. Octarine can connect to Ollama as an AI provider for Writing Assistant and Ask Octarine.

### Does Ollama keep AI requests local?

By default, Ollama runs on `localhost`, so prompts and model responses stay on your machine unless you configure Ollama differently.

### What Ollama URL should I enter in Octarine?

Use `http://localhost:11434` unless you changed Ollama's default server address.