Grux AI Download

How do I stop, update, move or uninstall Ollama on a Mac?

Ollama's macOS app registers as a login item when you install it. To stop it starting at login, open Settings, search for Login Items, find Ollama under Allow in the Background and switch it off; Ollama says it respects that across upgrades, unless you uninstall the app. A model stays in memory for 5 minutes after its last request, and ollama stop <model> unloads it at once. Models are stored in ~/.ollama/models, and the OLLAMA_MODELS environment variable points Ollama at a different folder. Ollama downloads updates by itself and applies them when you choose Restart to update. Uninstalling means deleting the app, the command line link and several folders, and the last of them, ~/.ollama, holds your models unless you moved them with OLLAMA_MODELS. This page follows Ollama's own documentation and covers the app, not a Homebrew install.

Download Grux AI 3.3.1, free, for macOS 14 or later on Apple silicon.

Grux AI 3.3.1 · sources read 2026-11-19 · published 2026-11-19

Stop Ollama starting when you log in

  1. Open Settings and search for Login Items.
  2. Find the Ollama entry under Allow in the Background.
  3. Click the slider to switch it off.

Ollama's FAQ says the macOS and Windows apps register as a login item during installation, and that the app respects this setting across upgrades unless you uninstall it. The FAQ does not describe how to quit a running Ollama, so this page does not either.

Unload a model, or keep it loaded

By default a model is kept in memory for 5 minutes, so the next request answers quickly. To free the memory at once:

ollama stop llama3.2

Through the API, the keep_alive parameter sets how long a model stays loaded: a duration such as "10m" or "24h", a number of seconds, any negative number to keep it loaded, or 0 to unload it right after the reply. The OLLAMA_KEEP_ALIVE environment variable sets the same thing for every model, and the API parameter overrides it.

Where models are stored, and how to move them

On a Mac, models are in ~/.ollama/models. ~/.ollama also holds Ollama's configuration, and ~/.ollama/logs holds the logs: app.log for the app and server.log for the server. To use a different directory, set the OLLAMA_MODELS environment variable to it. Ollama's macOS page warns that models can be tens to hundreds of GB in size.

When Ollama runs as the macOS app, the FAQ says to set environment variables with launchctl setenv and then restart the Ollama app. The path below is a placeholder for your own folder.

launchctl setenv OLLAMA_MODELS "/path/to/your/folder"
# then restart the Ollama app

Ollama's documentation does not say whether models you have already downloaded are moved for you when you change the setting, so this page does not say either.

Update Ollama

On macOS Ollama downloads updates by itself. Click the menu bar item, then "Restart to update", to apply one. You can also install an update by downloading the latest version from ollama.com/download.

Uninstall Ollama

Ollama's macOS page says that to fully remove Ollama you delete these files and folders. The last line, ~/.ollama, is where the same page says models and configuration live, so it deletes every model you downloaded, unless you moved them with OLLAMA_MODELS; skip it to keep them.

sudo rm -rf /Applications/Ollama.app
sudo rm /usr/local/bin/ollama
rm -rf ~/Library/Application\ Support/Ollama
rm -rf ~/Library/Saved\ Application\ State/com.electron.ollama.savedState
rm -rf ~/Library/Caches/com.electron.ollama/
rm -rf ~/Library/Caches/ollama
rm -rf ~/Library/WebKit/com.electron.ollama
rm -rf ~/.ollama

That page covers the Ollama app. It does not describe a Homebrew install, and neither does this page.

If you use Grux AI

Grux AI uses a local Ollama to run local models, so deleting ~/.ollama also removes the models Grux would have used. Which LLM can my Mac run says which models fit your memory before you download them again.

Where to check this

ClaimSource
Login item, how to switch it off, respected across upgradesOllama FAQ, How can I stop Ollama from starting when I login
Five minutes in memory, ollama stop, keep_alive and OLLAMA_KEEP_ALIVEOllama FAQ, keep a model loaded or unload it
Models in ~/.ollama/models, OLLAMA_MODELS, launchctl setenv, automatic updatesOllama FAQ
~/.ollama, the log files, the uninstall list, tens to hundreds of GBOllama documentation, macOS

Questions

How do I stop Ollama from starting at login on a Mac?
Open Settings, search for Login Items, find Ollama under Allow in the Background and click the slider to switch it off. Ollama's FAQ says it respects the setting across upgrades.
Where does Ollama store models on a Mac?
In ~/.ollama/models. Set the OLLAMA_MODELS environment variable to use a different directory; when Ollama runs as the macOS app, set it with launchctl setenv and restart the app.
How do I unload a model from memory?
Run ollama stop with the model's name. Otherwise a model is unloaded after 5 minutes by default, a time the OLLAMA_KEEP_ALIVE variable changes.
How do I uninstall Ollama and its models from a Mac?
Delete the app, the command line link and the folders Ollama's macOS page lists, ending with ~/.ollama, which holds your downloaded models unless you moved them with OLLAMA_MODELS. Skip that last one to keep them.
Download Grux AI 3.3.1 Read the source

Free, MIT licensed. macOS 14 or later, Apple silicon. 29.1 MB, signed and notarized by Apple. No account, no server, no subscription.