Quickstart
This guide will help you get to a working setup with Kubetorch using its base default settings. You will:
- Install the Python client with
piporuv - Configure a Kubernetes cluster with Kubetorch installed
Python Client Installation
Kubetorch provides a Python client for interacting with your cluster, and should be installed for both
local development and within your Docker images.
You can install it either with pip or uv, which offers faster resolution and
reproducible lockfiles.
pip install "kubetorch[client]"
Note
If you are running Kubetorch from a Mac, you should update rsync with: brew install rsync.
Mac devices ship with an older version of rsync that is missing modern features required by Kubetorch for code and data syncing.
Connect to a Kubernetes Cluster
You can install Kubetorch on your own Kubernetes cluster, or get started with our managed, serverless Kubetorch experience.
Kubetorch Serverless
Try out Kubetorch on managed, “serverless” compute. You’ll get quick access to a Kubernetes cluster with Kubetorch installed and ready to use.
Get StartedYour Own Cluster
Visit our Kubernetes Installation Guide for a walkthrough on how to get started with helm chart installation, as well as additional configuration options.
K8s InstallationIf you have any questions or need help, join our Slack community or reach out to the Runhouse team.