> ## Documentation Index
> Fetch the complete documentation index at: https://starforge.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Deployment options

> Four backends: local, node, kuberay, slurm

The console is one control-plane process. Where training containers run is a property of the **Fleet** a job is placed on, so one console can carry several backends at once (see [Fleets](/en/ops/fleets)); `FORGE_DEFAULT_FLEET_KIND` only says which one to seed at first start. User commands stay the same. There is no runtime fallback if the chosen backend is misconfigured.

## Comparison

|                          | **local**                         | **node**                                           | **kuberay**                             | **slurm**                                              |
| ------------------------ | --------------------------------- | -------------------------------------------------- | --------------------------------------- | ------------------------------------------------------ |
| Training vehicle         | Docker on the console host        | Docker on a remote node (`forgelet`)               | K8s RayJob (head + workers)             | slurmrestd allocation (Enroot/Pyxis or Apptainer)      |
| Scale                    | One node, 1–8 GPUs                | Many single-node jobs                              | Multi-node                              | Multi-node HPC                                         |
| Multi-node gang          | No                                | No (Phase 1; gang is later)                        | Yes                                     | Yes (`srun` + Ray)                                     |
| Heterogeneous multi-pool | No                                | No                                                 | Yes (worker groups)                     | Yes (hetjob + constraint)                              |
| Job code                 | Host directory mount              | Shared storage (`FORGE_STORAGE_ROOT`)              | Object storage + in-container bootstrap | Shared storage (`FORGE_STORAGE_ROOT`)                  |
| Playground               | Yes                               | No                                                 | Yes (vLLM Pod)                          | No                                                     |
| Extra deps               | Docker + NVIDIA Container Toolkit | Per-node daemon + shared disk                      | K8s + KubeRay + object storage          | slurmrestd (JWT) + container runtime                   |
| Where the console runs   | Same machine as the GPUs          | Anywhere that can reach agents and the shared disk | In-cluster or with a kube token         | Anywhere that can reach slurmrestd and the shared disk |

## How to choose

```mermaid theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}}
flowchart TD
    A["What compute do you have?"] --> B["One GPU server"]
    A --> C["A few bare-metal boxes, no K8s"]
    A --> D["Kubernetes"]
    A --> E["A Slurm site"]
    B --> L["local"]
    C --> AG["agent"]
    D --> K["kuberay"]
    E --> S["slurm"]
```

* **One box**: `local`. Fewest moving parts.
* **Two to eight machines, no K8s and no Slurm**: `agent`. Console no longer needs the Docker socket.
* **Existing Kubernetes**: `kuberay`. One RayCluster per job, K8s reclaims it.
* **Existing Slurm**: `slurm`. The platform is the submit and observe layer; the site scheduler still owns the queue.

## Pages

<CardGroup cols={2}>
  <Card title="Shared install" icon="download" href="/en/ops/install">
    Console process, database, object storage, reverse proxy
  </Card>

  <Card title="Fleets" icon="layers" href="/en/ops/fleets">
    Registering machines, join tokens, drain, liveness
  </Card>

  <Card title="local" icon="hard-drive" href="/en/ops/executor-local">
    Single-node Docker
  </Card>

  <Card title="node" icon="monitor" href="/en/ops/executor-node">
    Bare-metal machines running forgelet
  </Card>

  <Card title="kuberay" icon="boxes" href="/en/ops/executor-kuberay">
    Operator, RBAC, job packages
  </Card>

  <Card title="slurm" icon="network" href="/en/ops/executor-slurm">
    slurmrestd, runtime profiles, hetjob
  </Card>
</CardGroup>