Try it on your workstation

Two levels. Each says what it proves.

Get the code. The source is in the Apelora/apelora-fabric repository on GitHub. Clone it, then run the commands below from its top folder.

1. The designer only

npm install
npm run dev          # Studio at http://127.0.0.1:5173

Open the address. You get Apelora Studio with the sample application on the canvas. Everything runs in the browser against mock data. Nothing is sent anywhere and nothing is deployed: the header says Mock · Design only. Good for learning the canvas.

2. The whole stack, against a test cluster

This starts PostgreSQL, a disposable four-server Incus cluster (virtual machines standing in for physical servers), the control plane, the worker and both screens.

docker compose -f infra/dev/compose.yaml up -d       # PostgreSQL on 127.0.0.1:15432
bash scripts/test-cluster/up.sh                      # the test cluster (needs Lima: brew install lima)
source ~/.config/apelora/test-cluster/env            # its address and client certificates
cargo run -p apelora-control-plane                   # API at http://127.0.0.1:8080
cargo run -p apelora-deployment-worker               # carries out the jobs (second terminal)
npm run dev                                          # Studio at http://127.0.0.1:5173
npm run dev:console                                  # Fabric Console at http://127.0.0.1:5174

Run each long-lived command in its own terminal. Inside the repository, cargo finds the development database by itself.

Create your first API token (shown once):

cargo run -p apelora-control-plane -- token create --name me --role fabric-admin

Paste it into the Console's sign-in page. Details in sign in.

No database, no sign-in. DATABASE_URL= cargo run -p apelora-control-plane starts the API with mock data and no authentication. It only listens on the loopback address. Use it for a quick look, never on a shared machine.

What a test cluster can show you

Everything Apelora does with Incus: deployments, networking, volumes, and recovery when a server dies. It cannot show what depends on real hardware: disk and network speed, a management controller (BMC), or network boot. Try those on your own servers.

When you are done: bash scripts/test-cluster/down.sh removes the test cluster.