Ollama is a runtime for large language models (LLMs, neural networks that read and write text). It downloads open models and runs them on a Mac, Windows or Linux computer, or in Ollama's own cloud. The code is published on GitHub under the MIT license.
After installation, the command `ollama run gemma4` downloads a model and opens a chat with it. A REST API (an HTTP interface, here on the local port 11434) lets other programs send chat requests to the model, and official libraries exist for Python and JavaScript. The `ollama launch` command connects Ollama to coding agents such as Claude Code, Codex and OpenCode. Ollama states that prompts are never stored or used for training.
Running models locally is free and needs no subscription. Cloud models are paid per token (a token is a piece of a word), or through Pro, Max and Team plans that include monthly usage credits.