> ## 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.

# Sandbox0 sandboxes

> Run CostGraph workflows and CI jobs in Sandbox0 sandboxes: the key to create, its permissions, and its limits

Sandbox0 provides AI agent sandboxes so you can give every user their own cloud computer.

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 Sandbox0 on Integrations, and CostGraph starts sandboxes in your Sandbox0 account. You don't run anything yourself.

1. In CostGraph, go to **Integrations > Sandbox > Sandbox0**.

2. Enter these fields:

   * **API key** (`SANDBOX0_TOKEN`, required): A team-scoped Sandbox0 API key.

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 `sandbox0` 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

SANDBOX0\_TOKEN (or SANDBOX0\_API\_KEY) is a team-scoped Sandbox0 API key: "s0 auth login", `s0 team use <team-id>`, then `s0 apikey create --name ... --role developer --expires-in 30d --raw`.

Keys are team-scoped and carry a role, such as `developer`. Sandbox0 doesn't document what each role permits.

Sources: [sandbox0.ai: get started](https://sandbox0.ai/docs/get-started), [sandbox0.ai: pause resume](https://sandbox0.ai/docs/sandbox/pause-resume), [github.com: sandbox0](https://github.com/sandbox0-ai/sandbox0).

## Limits

* Soft TTL (ttl) auto-pauses, hard TTL (hard\_ttl) deletes state; both default 0 (disabled), so a sandbox can run as long as needed. ([source](https://sandbox0.ai/docs/get-started))
* Sandbox0 Cloud auto-pauses running sandboxes when unpaid usage reaches 10% of the last top-up (filesystem-only checkpoint; processes and memory lost). ([source](https://sandbox0.ai/docs/sandbox/pause-resume))
* vCPU/RAM caps, regions, architecture: not documented. ([source](https://sandbox0.ai/docs/get-started))

## 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.