Ollama
The easiest way to run a local model — one command, sensible defaults
Tools
In plain English
Like an app store for AI models that live on your own computer. You type the name of one and it downloads and runs, without you choosing any settings.
Technically
A wrapper around llama.cpp that adds a model registry, automatic GGUF download, prompt templating per model, GPU layer offload detection and an OpenAI-compatible HTTP server on port 11434. Models are described by Modelfiles, which pin a base model plus parameters and a system prompt. The tradeoff versus llama.cpp directly is control: Ollama picks quantization, context length and offload for you, and its defaults are conservative, so a tuned llama.cpp or vLLM setup will usually beat it on the same hardware. It is the right first stop and rarely the last one.