Ollama on a Mac: Requirements, Install, Where Files Live and How to Uninstall
Note: Some future links on this site may be affiliate links. We never claim hands-on testing; guides are researched from the public sources listed in each article.
How we wrote this: A summary of Ollama's official macOS page and FAQ (linked below), checked October 29, 2026. Requirements and file paths can change between releases.
Macs with Apple Silicon are one of the simplest ways to run AI models locally, because the CPU and GPU share one pool of memory. Here is what Ollama's documentation says about setting it up on macOS. (On a PC, see Ollama on Windows.)
Check the requirements
Ollama's macOS page lists:
- macOS Sonoma (version 14) or newer
- Apple M-series chip for CPU and GPU support. Intel (x86) Macs are CPU only.
On an Intel Mac, models still run, just more slowly. GPU acceleration on Apple Silicon goes through Apple's Metal API (see is Ollama using my GPU?).
Install
The docs recommend mounting the ollama.dmg download and dragging the Ollama app into the system-wide Applications folder. On first start, the app checks whether the ollama command-line tool is on your PATH; if it isn't, it asks permission to create a link in /usr/local/bin.
Then open Terminal and try a model, as in our Ollama beginner guide.
Installing somewhere else: put the app where you want it, make sure the CLI (Ollama.app/Contents/Resources/ollama, or a symlink to it) is on your PATH, and decline the "Move to Applications?" prompt on first start.
Plan for disk space
The docs warn that models can be tens to hundreds of GB. If your home folder is short on space, you can store models elsewhere by setting the OLLAMA_MODELS environment variable to another folder (the FAQ's default Mac location is ~/.ollama/models).
Where Ollama keeps things on a Mac
| Path | What's there |
|---|---|
~/.ollama |
Models and configuration |
~/.ollama/logs/app.log |
Recent logs from the app |
~/.ollama/logs/server.log |
Recent server logs |
<install location>/Ollama.app/Contents/Resources/ollama |
The CLI binary |
The server log is the first place to look when a model won't load.
Setting environment variables
The FAQ says that when Ollama runs as the macOS app, you set environment variables with launchctl, then restart the app. For example, to change the address the server listens on:
launchctl setenv OLLAMA_HOST "0.0.0.0:11434"
Think before you do that one: 0.0.0.0 makes the server reachable from other devices on your network, not just your Mac. The same method works for OLLAMA_MODELS and the other settings in the FAQ.
Updates
According to the FAQ, Ollama on macOS downloads updates automatically. Click the menu bar icon and choose Restart to update to apply one, or download the latest version manually from Ollama's site.
Uninstall completely
The macOS page lists every location to remove for a full uninstall:
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
The last line deletes your downloaded models, so copy anything you want to keep first. Note that the docs put two of those paths in quotes with a ~ inside; in most shells a quoted ~ isn't expanded, so if those folders remain afterward, delete them from Finder or re-run those two lines without the quotes around the ~.