Ollama Operator

Ollama Operator is a Kubernetes operator for running large language models on a cluster, powered by Ollama. It extends the Ollama experience to any Kubernetes environment, whether edge hardware or a cloud provider, using the same specification everywhere. A model is deployed by declaring a custom resource of kind Model, which the site shows can be as short as a few lines of YAML naming the model image. Users who prefer to avoid YAML can use the companion kollama command line tool, either as a standalone binary or as a kubectl plugin, with a command such as deploy followed by a model name. Once deployed, the model can be queried through an OpenAI compatible chat completions endpoint, so existing code and clients that use that request and response format work without changes. The site also states that it works with LangChain for function calling, agents and knowledge base retrieval. The documentation is available in English and Simplified Chinese, and the project is hosted on GitHub. It is aimed at platform and infrastructure engineers who want to serve open models on Kubernetes with GitOps and CI and CD workflows. The site lists no prices. Costs depend on the user's own cluster infrastructure.