Prodigy

Your applications. Your machines. One orchestrator.
Provision machines, deploy applications, and manage their lifecycle together.

  Documentation · Examples

Early-stage software for evaluation. Runtime integration required; OCI images need adaptation.

Run → inspect → update → remove
Watch the walkthrough · Read the transcript

Run your first service

The example brings a real HTTP application, two versions, and matching runtime tools. Explore a disposable one-machine cluster without a cloud account or writing application code.

Prepare the evaluation bundle, open its directory, and run:

./try-prodigy

Keep it running. In another terminal, use ./try-prodigy status to inspect the service and ./try-prodigy update to change its response from v1 to v2. Press Ctrl-C in the first terminal to remove the cluster; on macOS, the launcher also stops its guest.

The candidate currently requires a source build and a supported Linux guest; Apple Silicon uses Apple Containers. Requirements and availability · Complete tutorial.

Why Prodigy

  Manage the machines and the applications.
Provision capacity, place workloads, and manage their lifecycle through one system—across machines you own and infrastructure-provider adapters.

  Give applications the context they need.
Deliver configuration, topology, credentials, and resource changes directly to workloads. The included example already implements the required runtime integration.

  Keep deployment and traffic in step.
Coordinate where instances run, how requests reach them, and when they start, update, and stop.

How Prodigy works · Compare with Kubernetes and Nomad

Make it yours

  Build an application →
Start with the example source, then integrate your own workload.

  Use your machines →
Bring hardware you operate into the cluster.

  Create cloud capacity →
Provision machines through Prodigy's infrastructure-provider adapters.

---

Current status · Contributing · Support · Security · Apache-2.0