> ## 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.

# Isorun sandboxes

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

Isorun is an execution runtime for AI agents that runs isolated Linux microVMs, each with its own kernel, booting in milliseconds.

The following table shows what CostGraph runs on this provider.

| Workflows | CI runners | Connect with |
| - | - | - |
| Yes | Yes | Integrations or your own dispatcher |

## Connect

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

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

2. Enter these fields:

   * **API key** (`ISORUN_API_KEY`, required): An Isorun API key from the Isorun dashboard.

3. Click **Connect**.

For CI runners, CostGraph sets CPU and memory and boots `providerTemplate` as the image.

### Your own dispatcher

To keep the key out of CostGraph or reach a private network, run your own dispatcher instead.
Set `provider` to `isorun` 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

ISORUN\_API\_KEY is an Isorun API key (isorun\_live\_\<region>*\<id>*\<hmac>). Create it in the dashboard at app.isorun.ai. Each key belongs to your organization and that organization's region (US or EU).

keys are not scoped: one API key has full access to its organization (in its region). The docs describe no scopes, permission levels or roles. Source: [https://docs.isorun.ai/getting-started/authentication](https://docs.isorun.ai/getting-started/authentication)

Sources: [docs.isorun.ai: home](https://docs.isorun.ai/), [docs.isorun.ai: authentication](https://docs.isorun.ai/getting-started/authentication), [docs.isorun.ai: lifecycle](https://docs.isorun.ai/sandboxes/lifecycle), [docs.isorun.ai: troubleshooting](https://docs.isorun.ai/troubleshooting), [docs.isorun.ai: pricing](https://docs.isorun.ai/getting-started/pricing).

## Limits

* Idle timeout: timeoutSec defaults to 300 s, and the sandbox is destroyed when it fires. Extend it with setTimeout(seconds); 0 disables it. No maximum lifetime is documented. ([source](https://docs.isorun.ai/sandboxes/lifecycle))
* Default size is 1 vCPU, 1 GiB RAM and 4 GiB scratch disk. No maximum vCPU or memory is documented. ([source](https://docs.isorun.ai/sandboxes/lifecycle))
* exec has a default timeout of 30 seconds; pass a longer timeout or use execStream for long work. Concurrent sandboxes count against your limits (429), but the number is not published. ([source](https://docs.isorun.ai/troubleshooting))
* Regions are US and EU, fixed per organization and taken from the key. The docs do not state the CPU architecture. ([source](https://docs.isorun.ai/getting-started/authentication))

## 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.