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

# HopX sandboxes

> Run CostGraph workflows and CI jobs in HopX sandboxes: the key to create, its permissions, and its limits

HopX runs secure, isolated cloud sandboxes with dedicated kernels that AI agents use to execute code. The dedicated kernels provide hardware-level isolation.

The following table shows what CostGraph runs on this provider.

| Workflows | CI runners | Connect with |
| - | - | - |
| Yes | No | Integrations or your own dispatcher |

CI runners: no. AutoPilot compares runners by price, and this provider publishes no sandbox price.

## Connect

Connect HopX on Integrations, and CostGraph starts sandboxes in your HopX account. You don't run anything yourself.

1. In CostGraph, go to **Integrations > Sandbox > HopX**.

2. Enter these fields:

   * **API key** (`HOPX_API_KEY`, required): A HopX API key from the HopX console.

3. Click **Connect**.

### Your own dispatcher

To keep the key out of CostGraph or reach a private network, run your own dispatcher instead.
Set `provider` to `hopx` and the preceding fields as environment variables, as described in [Run your own dispatcher](https://docs.costgraph.ai/costgraph/workflows/managed-sandboxes.md#run-your-own-dispatcher).

## Key and permissions

HOPX\_API\_KEY is a HopX API key in the format hopx\_live\_\<keyId>.\<secret>. Get it from the dashboard at console.hopx.dev. The dashboard shows it only when you create it.

HopX names no key scopes. Its docs mention limited keys only where your organization supports them.

Sources: [docs.hopx.ai: api key](https://docs.hopx.ai/api-key.md), [docs.hopx.ai: authentication](https://docs.hopx.ai/api/authentication.md), [docs.hopx.ai: timeout](https://docs.hopx.ai/core-concepts/sandboxes/timeout.md), [docs.hopx.ai: create sandbox](https://docs.hopx.ai/api/control-plane/create-sandbox.md), [docs.hopx.ai: build template](https://docs.hopx.ai/api/control-plane/build-template.md).

## Limits

* timeout\_seconds is an auto-kill timeout. When it expires the sandbox is destroyed and its data is lost. It can be extended with set\_timeout. If no timeout is given, the sandbox runs indefinitely. No maximum is documented. ([source](https://docs.hopx.ai/core-concepts/sandboxes/timeout.md))
* Resources come only from the template. You cannot set them per sandbox. Template memory is 512-64000 MB, disk 1-250 GB, and the vCPU range depends on your plan. ([source](https://docs.hopx.ai/api/control-plane/build-template.md))
* You can pick a region, for example us-east or eu-west. The docs do not state the CPU architecture. ([source](https://docs.hopx.ai/api/control-plane/create-sandbox.md))
* Rate limits: 20 sandbox creations per minute, 100 control-plane requests per minute and 300 VM agent requests per minute. ([source](https://docs.hopx.ai/api/concepts/rate-limits.md))

## Related

* [Supported sandbox providers](https://docs.costgraph.ai/costgraph/workflows/sandbox-providers.md)
* [Run workflows on a managed sandbox provider](https://docs.costgraph.ai/costgraph/workflows/managed-sandboxes.md)
* [Run CI jobs in sandboxes](https://docs.costgraph.ai/costgraph/autopilot/ci-sandboxes.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.