Start here · Step 5 of 7

1. What is a local LLM
2. Hardware
3. Words
4. Pick a model
5. Install
6. First chat
7. Next

Step 5 of 7 · about 8 minutes

Install

One runner is enough for a first chat. Do not install all three today.

Ollama

This is the first installer for most people: one official download, a short command to pull a model, and a local API. Official page: ollama.com/download. Official docs: docs.ollama.com/quickstart.

Install commands published there, checked 19 Nov 2026: on macOS and Linux, curl -fsSL https://ollama.com/install.sh | sh. On Windows PowerShell, irm https://ollama.com/install.ps1 | iex. macOS and Windows also have official app downloads on that page. Those commands come from Ollama, not from a store.

LM Studio

Pick this if you want a desktop chat window and a model browser, without a terminal. Official site: lmstudio.ai. Official downloads: lmstudio.ai/download. I am not copying a version number off that page into this lesson, because it changes.

llama.cpp

Pick this if you want the GGUF engine itself. Official repository: github.com/ggml-org/llama.cpp. The v1.0.0 announcement (13 Nov 2026) lists official distribution channels, including GitHub artifacts, Docker, Debian, Homebrew, conda-forge, MacPorts, and Winget. I am not inventing an install one-liner that is not on that repo.

Some official Ollama tags are cloud tags. They end in :cloud on Ollama’s own download page. This site is about running models on a machine you have. Cloud tags are out of scope for the first chat.