Self-Managed Installations
JuliaHub can be deployed as a self-managed installation in your own infrastructure. Choose the installation method that best fits your environment:
Dyad AI is not enabled by default on self-managed (on-premises) installations, regardless of the installation method below. Enabling it requires the appropriate license entitlement and additional configuration, so it is not a self-serve toggle. If you would like to use Dyad AI on your installation, contact JuliaHub support to confirm your entitlement and the supported configuration.
Available
AWS (Helm) — Deploy the full JuliaHub platform on an Amazon Elastic Kubernetes Service (EKS) cluster using Helm. Includes compute support for batch jobs and applications. Terraform modules are not yet provided, so the guide enumerates the AWS resources to provision and how to wire them into the chart.
Azure (Helm) — Deploy the full JuliaHub platform on an Azure Kubernetes Service (AKS) cluster using Helm. Includes compute support for batch jobs and applications.
Embedded Cluster — Deploy JuliaHub on a single Linux machine with no existing Kubernetes cluster required. Currently provides the package server only (compute support is planned for a future release).
Coming Soon
- GCP (Helm) — Deploy the full JuliaHub platform on Google Kubernetes Engine (GKE) using Helm.