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

# Quickstart

> Install the CLI, sign in, and create your first sandbox.

This quickstart covers installing the CLI, signing in, creating a project, and the first
sandbox: create, inspect, delete.

## 1. Sign in and choose a project

Install the CLI, point it at your deployment, and authenticate:

```sh theme={null}
curl -fsSL https://hiloop.ai/install.sh | sh
export HILOOP_API_URL=https://api.hiloop.ai
hiloop login
hiloop whoami
```

On-prem operators supply their own API URL and trust configuration. Create a project once, then use
its slug for the rest of the guide:

```sh theme={null}
hiloop projects create demo --name "Sandbox demo"
export HILOOP_PROJECT=demo
```

If `demo` already exists, keep using it.

## 2. Create a sandbox

Create a sandbox, bounded by a short TTL for a first run. The CLI waits until it is running:

```sh theme={null}
hiloop sandbox create demo-box --ttl 900
```

With no `--image`, the sandbox starts from the
[platform default image](/sandboxes/images#the-platform-default-image), a small base with Python
and Node preinstalled. To bring your own environment, pass `--image <ref>` with any OCI image; the
sandbox runs the image's own entrypoint, so pick an image that keeps running (a server, a shell
held open) rather than one that exits immediately.

Inspect it, then clean up:

```sh theme={null}
hiloop sandbox get demo-box
hiloop sandbox delete demo-box
```

If your deployment has the sandbox SSH endpoint enabled, `hiloop sandbox ssh demo-box` also opens a
shell in it, with stock OpenSSH and no key or `ssh` configuration from you: the credential you
signed in with authorizes the connection. That endpoint is off by default, and on the hosted service
it is currently enabled only in pre-release environments, so a connect may be refused with
`unsupported_capability`.

`hiloop sandbox exec demo-box -- <command>` runs one buffered command in the sandbox and returns its
real exit code; [what's live today](/sandboxes/overview#whats-live-today) tracks the current status
per verb. `hiloop run` executes a command locally with full telemetry capture.

Next: [run a sandbox](/guides/running-sandboxes) for the full verb tour, and
[snapshots and branching](/concepts/workspaces) for the persistence model.