Ollama installation tutorial: Install on Windows, macOS, and Linux

Windows

1. Download the Windows installer (OllamaSetup.exe) from ollama.com.
2. Run the installer. Administrator privileges aren't required, and Ollama installs in your user folder by default.
3. After installation, Ollama runs in the background. The Ollama icon appears in the system tray in the lower-right corner, and the Ollama desktop app opens.

Open PowerShell or Command Prompt to verify:

ollama --version

If you're using an NVIDIA GPU, update your graphics driver to a recent version first. Ollama automatically detects and uses the GPU.

macOS

1. Download the macOS version from ollama.com.
2. Unzip it, drag Ollama into the Applications folder, and open it.
3. The first time you open it, you'll be asked whether to install the ollama command-line tool. Allow it.

Apple Silicon (M-series) Macs automatically use GPU acceleration.

Linux

The official one-line installation script is:

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

The script will:

• Install the ollama executable
• Create the ollama system user
• Create and start the ollama.service systemd service
• Detect NVIDIA / AMD GPUs and configure the related components

Check the service status:

systemctl status ollama
ollama --version

View the service logs (useful for debugging):

journalctl -u ollama -f

If you prefer to inspect the script before running it, download it first with curl -fsSL https://ollama.com/install.sh -o install.sh, review it, and then run it.

Install with Docker

To run Ollama in a container, see Run Ollama with Docker.

Verify that the installation succeeded

ollama run llama3.2 "Introduce yourself in one sentence"

The first run downloads the model (about 2 GB). If you see a response from the model, the installation succeeded.

Check whether the GPU is being used

After the model loads, run this in another terminal:

ollama ps

NAME               ID              SIZE      PROCESSOR    CONTEXT    RUNNER      UNTIL
llama3.2:latest    a80c4f17acd5    3.5 GB    100% GPU     4096       llamacpp    4 minutes from now

The PROCESSOR field:

If it should be using the GPU but shows 100% CPU, see GPU and performance tuning.

Where the Ollama service runs

Ollama consists of two parts:

• Service (server): Runs in the background, loads models, and provides an API. By default, it listens on 127.0.0.1:11434.
• Command-line tool (CLI): The ollama command you enter, which calls the service's API.

The apps on Windows and macOS start the service automatically; on Linux, systemd manages it. You can also start it manually in the foreground (for example, when debugging):

ollama serve

If a service is already running, you'll see an address already in use error. That's normal.

Check whether the service is running:

curl http://localhost:11434

A response of Ollama is running means it is working.

Update

• Windows / macOS: The app checks for updates automatically. Restart it to apply them.
• Linux: Run the installation script again to upgrade.

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

Uninstall

• Windows: Uninstall it under Settings → Apps. The model files are in C:\Users\YourUsername\.ollama; delete them manually.

• macOS: Quit Ollama, delete it from Applications, and delete the ~/.ollama folder.

• Linux:

  sudo systemctl stop ollama
  sudo systemctl disable ollama
  sudo rm /etc/systemd/system/ollama.service
  sudo rm $(which ollama)
  sudo rm -rf /usr/local/lib/ollama   # GPU libraries (if present)
  sudo rm -r /usr/share/ollama
  sudo userdel ollama
  sudo groupdel ollama

Model files usually take up a lot of space, so remember to delete them when uninstalling.

Updated October 12, 2026