Ollama is the easiest way to run AI models locally. In just a few minutes, you'll have a model running directly on your machine — no cloud, no API keys, no friction.
Running AI models locally used to require complex toolchains, GPU drivers, and hours of setup. Ollama changes that. It gives you a simple, unified way to run modern large language models — including Llama 3, Mistral, and many others — directly on your machine with a single command.
Installing Ollama is your first real step into local‑first AI: fast, private, offline intelligence that lives on your hardware, not in the cloud. Choose your operating system below and follow the quick setup.
On macOS, Ollama integrates cleanly with the system and supports both Intel and Apple Silicon (M1/M2/M3) chips. If you’re on Apple Silicon, you’ll benefit from excellent on‑device acceleration.
Install using Homebrew:
Or download the official installer:
Windows support is now fully stable and includes GPU acceleration on supported NVIDIA hardware. The installer handles everything for you.
Download the official installer:
After installation, restart your terminal or PowerShell window.
Linux users can install Ollama using the official script. It works across major distributions including Ubuntu, Debian, Fedora, and Arch.
For advanced configuration, see the Ollama GitHub repository.
Confirm that Ollama is installed correctly by checking the version:
You should see something like:
If you see a version number, your system is ready to run local models.
Let’s run Llama 3, one of the most capable open models available today. Ollama will automatically download the model the first time you run it.
If you see a prompt and the model responds, congratulations — you’re officially running local AI on your machine. No cloud. No API keys. No data leaving your device.
You can explore more models in the Ollama Model Library.
“ollama: command not found”
Model download is slow
GPU not detected
Next Step
Download Your First Models →