Getting started

Adopt Kubeara by choosing how the control panel runs, connecting at least one Ubuntu server, and deploying a service from the catalog. Logs stream until the job completes.

Adoption path

1. Choose Cloud or self-hosted
2. Create an account (Cloud) or open the local dashboard (self-hosted)
3. Connect a server with SSH
4. Deploy a catalog service
5. Watch logs and verify the endpoint
6. Optionally connect MCP or the HTTP API

1. Choose a deployment approach

2. Connect a server

In the dashboard go to Servers and add:

• A display name
• Host (IP or DNS)
• Username
• SSH credentials (key or password, as your host requires)

The server must be reachable from the control panel. After connect, Kubeara validates the host and installs what the agent needs. You do not install Docker, open a Kubeara inbound port, or write nginx by hand.

Details: Servers and Requirements.

3. Deploy a service

Open Services, pick a stack (for example PostgreSQL or Grafana), choose the server, and click Deploy.

Kubeara:

• Pulls and runs the stack
• Provisions TLS
• Configures health checks and volumes
• Streams job logs until completion

If a step fails, the log output is the first place to look. Details: Services and Deployments.

4. Verify

• Use the assigned domain or the host plus the stack’s default port
• Confirm health checks are green
• For GPU models, check VRAM and utilization before pulling a larger model

5. Optional: MCP

• MCP — Claude, Cursor, ChatGPT, VS Code

Quick links