Connect Your Machines
Install a managed cluster over SSH, then add, inspect, or remove worker nodes as capacity changes.
You bring a machine you control; Studio turns it into a managed cluster you can deploy onto. The machine can be a public-cloud VM or a bare-metal box — all Studio needs is SSH access and outbound reachability.
Initial setup turns the first machine into the cluster's control-plane node. After it is ready, the Nodes page can join another reachable machine or merge an eligible standalone cluster as a worker. The control-plane node cannot be removed as a worker.
Requirements
- A reachable machine — its host address and SSH port.
- An SSH login — a username plus a key. Studio can generate a fresh key pair for you, or you can paste an existing private key.
- Outbound access on ports 80 and 443 so the cluster can obtain and serve TLS certificates for your deployments.
- Optionally, a machine with a GPU if you plan to run accelerated workloads — the wizard can install GPU drivers for you.
Setup Wizard
The wizard walks you through installation step by step. Progress and logs stream live, and a failed run can be retried without re-entering the form.
Choose the machine type. Provide the host, SSH port, and login user for the machine you want to bring in.
Set up the connection. Let Studio generate an SSH key pair, or paste your own existing private key.
Run preflight checks. Studio verifies that SSH connects, that ports 80 and 443 are reachable from outside, and whether the machine already has a GPU or an existing cluster on it.
Name the cluster and choose options. Pick a cluster name and toggle optional add-ons: GPU drivers, a monitoring stack, and DNS handling.
Submit and install. Studio connects over SSH and installs the cluster for you. Watch progress and logs in real time, and retry on failure without re-filling anything.
Add and manage worker nodes
Open the cluster and select Nodes to review live node health and recorded capacity. To expand the cluster:
Worker nodes can be removed from the same page. Removal queues a drain and cleanup job; the control-plane node is protected from this action.
Cluster Domains & Delegated Subdomains
Every cluster is associated with a workspace domain, and each deployment you create gets a subdomain beneath it automatically (see Deploying Services). You have three ways to supply that domain:
- A free Studio-managed subdomain — the quickest path, handled entirely for you.
- Your own DNS provider — Studio installs ExternalDNS on the cluster for supported providers and manages records for you.
- A manual (DIY) domain — you point the wildcard record at the cluster's public IP yourself.
See the DNS Service page for the full breakdown and the list of supported providers.
Managing & Cleaning Up a Cluster
From the deploy area you can see cluster status, nodes, capacity, and running deployments. Remove unneeded worker nodes individually when shrinking capacity. Deleting the cluster queues full cluster cleanup and releases its registered machines after the job succeeds.
Admins can cap how many clusters and domains each workspace may create. If you hit a limit, remove an unused cluster or ask your Admin to raise the cap.
Quick Start: From One Server to Your First App
One complete walkthrough — onboard a machine, carve it into compute, list it, rent it, set up DNS, and deploy an app. Every step has a screenshot.
Compute Marketplace
Deliver virtual clusters as a provider and publish provisioned capacity for future renters.