Draft:Ollama

Ollama is an open source^[1] (Ollama is under the MIT license.)software framework^[2] for managing and running large language models (LLM) locally or in the cloud. It offers a command-line interface (CLI) for downloading, installing, using and removing models. By providing a uniform interface to the models it supports, it enables users to use models whose details they are not familiar with allowing them to run AIs locally on computer hardware. Jeffrey Morgan ^[3] and Michael Chiang ^[4] founded Ollama in 2021.

At the core of Ollama is a local web server that provides a JSON REST API for managing models and interacting with them. The ollama program starts and manages that server and uses the REST API to provide a CLI for managing models and interacting with them.^[5]

On its own, Ollama's web server does not have a graphical user interface. For this purpose, the Open WebUI can be used.^[5]

Ollama has its own website ^[6]. This is where its subscription service comes into play. In addition, the website helps users to download the vast amount of AIs through the terminal.

Ollama is primarily written in Go. The software integrates inference components derived from llama.cpp, which are implemented largely in C and C++ for efficient execution of large language models on local hardware.^[7] Commonly, run on a terminal.