---
title: How Kapten works
description: The objects Kapten is built on, from the organization down to the components it deploys, and how they relate to each other.
sidebar:
  label: Overview
---

Kapten runs your workloads on Kubernetes clusters in your own cloud account. Its control plane, the **Engine**, stores what you want to run and drives an agent installed on each of your clusters to make it happen.

```text
Organization
├── Clusters            Kubernetes clusters you connected
├── Integrations        GitHub installations
└── Projects
    └── Environments    development, staging or production, each on one cluster
        ├── Components  applications, jobs and databases
        ├── Variables and secrets
        └── Pipelines
```

- An **[organization](/concepts/organizations-and-roles)** owns everything and decides who can do what through roles.
- A **[cluster](/concepts/clusters)** is a Kubernetes cluster that Kapten deploys to.
- A **[project](/concepts/projects)** groups the services that ship together.
- An **[environment](/concepts/environments)** is one deployment of a project, such as `staging` or `production`, on one cluster.
- A **[component](/concepts/components)** is something that runs in an environment: an application, a job or a database.
- **[Variables and secrets](/concepts/variables-and-secrets)** configure components, per project, environment or component.
- A **[pipeline](/concepts/pipelines)** deploys several components of an environment in ordered stages.

Every deployment of a component is recorded as a **run**, with its status, its timeline and its logs, so you can see what was deployed, when, and why it failed.
