ollama-helm
ollama-helm is a community Helm chart that deploys Ollama on Kubernetes, with configuration for CPU-only or GPU-accelerated clusters and for pulling or creating models automatically at startup.
ollama-helm is a Helm chart for installing and running Ollama, a tool for serving large language models locally, on a Kubernetes cluster. It supports Kubernetes 1.16 and later for CPU-only operation, and 1.26 and later for stable GPU acceleration on NVIDIA or AMD hardware, though not every GPU, especially some AMD ones, is currently supported by Ollama itself.
Values configuration can enable GPU support and choose NVIDIA or AMD, list models to pull automatically when the container starts, or define a new model from a template referencing an existing one with custom parameters. An example configuration also shows exposing Ollama's API through an ingress host.
The chart is distributed from the maintainers' own Helm registry rather than GitHub Pages, and Ollama itself can be reached afterward through its REST API, its official JavaScript and Python client libraries, or LangChain's JavaScript and Python integrations.