Documentation

Get started in 60 seconds.

Pick a path. Each one finishes with a working model on your machine.

Install

Get the CLI on macOS, Linux, or Windows.

Run a model

Pull, run, chat. The basics.

API reference

Drop-in OpenAI compatibility.

Cloud

Run models too big for your laptop.

Install

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

Run a model

Pull a model, then run it interactively.

ollama pull llama4:scout
ollama run llama4:scout

API

Send chat requests to your local server.

curl http://localhost:11434/api/chat -d '{
  "model": "llama4:scout",
  "messages": [{ "role": "user", "content": "Why is the sky blue?" }]
}'

Cloud

For models that don't fit on your laptop, run on Ollama Cloud — same CLI, same API, just --cloud.

ollama run --cloud kimi-k2:1t