Ollama

Ollama is an open-source program that lets you download and run AI models on your own computer in very few steps. It's for having an assistant that runs on your machine, without sending your data to the internet, and for letting other apps use those models as if they were a cloud service. It works on Windows, Mac and Linux.

Ollama was born to be used from the terminal, that text window where you type commands: you type something like "ollama run" followed by the model's name and start chatting. Today it also has a desktop app, but its strong point is still something else: it opens a "door" on your computer (what techies call an API) through which other programs can ask the model for answers. What usually confuses people is thinking Ollama is an AI. It isn't: it's the program that loads and runs the models, like DeepSeek, Qwen, Gemma or Llama. And it doesn't work miracles with your hardware: if the model doesn't fit in your memory, it'll be slow all the same.

Ollama is one of the four tools I use for local AI, along with LM Studio, llama.cpp and Unsloth. I call these pieces harnesses: whatever makes a model usable. In "Ollama, LM Studio, llama.cpp and Unsloth: which one I use for what" I explain what I use each one for, how I ended up using all four rather than just one, and what I'd recommend depending on what you want to do.