What Ollama does
Ollama runs open AI models on your own computer and also offers cloud models through the same app, CLI, and API. You can connect it to coding agents and desktop apps, and local use is always free.
Ollama is a tool for using open-weight AI models. You install it on macOS, Windows, or Linux, or run the official Docker image, and then download and run models from its library on your own hardware. Models for coding, vision, embeddings, and reasoning are available, and anything you run locally stays on your machine.
Ollama also offers cloud models for people who do not have a strong local GPU. The same app, command line, and API can send requests to models hosted by Ollama in the US, Europe, and Singapore. Ollama says prompts are not tracked or used for training.
A main use today is connecting open models to existing tools. One command launches coding agents such as Claude Code, Codex, or OpenCode with an Ollama model, and the desktop app can connect to apps like Claude. Developers can use the Ollama REST API, OpenAI-compatible and Anthropic-compatible endpoints, and official Python and JavaScript libraries, either against a local server or the hosted API.
Running models on your own computer is free and unlimited. The Free plan adds starter credits for a small set of cloud models. Pro costs $20 per month (or $200 per year) with $60 of monthly usage credits, Max costs $100 per month with $300 of credits, and a Team plan starts at $500 per month. Pay-as-you-go credits are priced per million tokens for each model.
Ollama is made by Ollama Inc. The core project is open source under the MIT license on GitHub, and development happens in the open with a Discord and Reddit community. The website says more than 9 million developers use it.
Best for
Developers and technical users who want to run open AI models privately on their own hardware or use hosted open models with their coding agents.
Not for
People who want a finished chat assistant with no setup, or teams that need a specific closed model such as GPT or Claude.
Questions and answers
Is Ollama free?
Running models on your own computer is always free and unlimited. The Free plan also includes starter credits for some cloud models.
How much do the paid plans cost?
Pro is $20 per month or $200 per year with $60 of usage credits each month. Max is $100 per month with $300 of credits, and Team starts at $500 per month.
Which operating systems does it support?
Ollama runs on macOS, Windows, and Linux, and there is an official Docker image. The macOS app requires macOS 14 Sonoma or later.
Does Ollama have an API?
Yes. Ollama has its own REST API plus OpenAI-compatible and Anthropic-compatible endpoints, for a local server at localhost:11434 or for the hosted API at ollama.com.
Is my data used for training?
Ollama says your prompts are never tracked or trained on, and nothing you run locally leaves your machine.