TopolearnEvery other lesson assumes a working install with the server listening on its default port.
Download Ollama from the official site, ollama.com/download.
- macOS and Windows: install the desktop app like any other program. It starts the Ollama server in the background and shows an icon in the menu bar or system tray. The app also includes a chat window and settings.
- Linux: run the official install script in a terminal:
curl -fsSL https://ollama.com/install.sh | sh. It installs theollamacommand and sets it up as a background service (on most distributions with systemd, the system that starts services at boot).
Either way you get the `ollama` command for the terminal and a server listening on your machine at http://localhost:11434. "localhost" means "this computer", and 11434 is the port, a numbered door that programs use to reach the server.
Check it works:
ollama --version
curl http://localhost:11434The second command should reply "Ollama is running". If it doesn't, start the app (macOS, Windows) or run ollama serve in a terminal (Linux, when the service isn't running).
Updating: macOS and Windows download updates automatically and ask you to restart. On Linux, run the install script again.
Downloaded models are large (often several gigabytes each) and stored in your home folder by default (~/.ollama/models on macOS and Windows). Make sure you have disk space; a later lesson shows how to move them.
Example
# macOS / Windows: install the app from https://ollama.com/download
# Linux:
curl -fsSL https://ollama.com/install.sh | sh
ollama --version # ollama version is 0.40.x
curl http://localhost:11434 # Ollama is running
# If the server is not running (e.g. Linux without the service):
ollama serveRemember
- App on macOS/Windows, install script on Linux
- The server listens on http://localhost:11434
ollama --versionandcurl localhost:11434confirm it works
Watch out for
- Running
ollama servewhile the app or service is already running, which fails because the port is already in use.