Ollama
The simplest way to download and run open models on your own computer: one command to chat, plus a local API other apps can use.
Ollama is a free, open-source tool that downloads open models and runs them on your own machine. You install it on macOS, Windows or Linux, then run a single command such as ollama run gemma4 to start chatting. Its model library at ollama.com/library lists ready-to-run versions of many open models, so you do not have to hunt for files or convert formats.
Ollama also runs a local REST API, by default on port 11434, so other programs can use your model. Official Python and JavaScript libraries are available, and an official Docker image exists for servers. A launcher can connect coding assistants such as Claude Code or Codex, and personal assistants, to a model served by Ollama. According to its README, you can use the latest model releases locally or in the cloud.
For most beginners it is the best first tool to try: installation takes minutes and the defaults work. Heavier use, such as serving many users at once, is better handled by an engine built for it, such as vLLM or SGLang.
Good for
- Trying an open model on a laptop or desktop for the first time
- Giving your own app or script a local model through a simple API
- Connecting coding assistants and chat front-ends such as Open WebUI to a local model
- Running on a home server with the official Docker image
Watch out for
- Library models are repackaged and are often quantized to fit consumer hardware, so check the tag before comparing quality with the original release
- Ollama can also run models in its cloud, so confirm whether a model runs locally before sending sensitive data
- It is aimed at local and small-team use; serving many users at once calls for a server engine