What it is
Ollama downloads and runs open models such as Gemma, gpt-oss, DeepSeek and GLM on Mac, Windows and Linux. You use it from the command line, its desktop app, or a local API that other apps and coding tools can call. It also offers cloud models that run on Ollama's servers when a model is too big for your hardware, and Ollama says prompts and responses are never logged or trained on.
Running models on your own computer is free and unlimited, and the project is open source. Cloud use is pay as you go, or Pro at $20 a month ($60 of usage credits) and Max at $100. Local speed depends on your hardware: large models need lots of memory and a capable GPU or Apple silicon.
Best for
- Running open models privately on your own machine
- Developers who want a local, OpenAI-style API for testing
- Using larger open models in the cloud when your hardware is too small