---
title: "Runners"
description: "Where Bugbearer runs test cases: cloud sandboxes or local Docker."
---

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

# Runners

A runner is the place where one test case runs in a real browser. Each case gets its own isolated environment. The runner records video, runs the checks, and writes a report.

You pick a runner per app with **Runs on** in the app settings. Workspace settings list the runners under **Runners**.

| Runner | What it is | Reaches |
| --- | --- | --- |
| Bugbearer cloud | One E2B sandbox per case. Nothing to install. | Public URLs. Tunnels: coming soon. Not a private network. |
| Local | One Docker container per case, on the machine or runner host that Bugbearer uses. | Apps that Bugbearer boots from a repo, on a private Docker network. |
| Your server | Docker on your own host that connects out to Bugbearer. | Coming soon |

## Which one to use

- **URL app:** Bugbearer cloud. This is the default.
- **Repo app:** Local. Bugbearer boots the app on a private Docker network, which a cloud sandbox cannot reach. A run on the cloud runner fails with a message that tells you to set **Runs on** to Local.

## Limits

- A cloud sandbox restricts outgoing traffic to an allowlist of hosts.
- Each case container or sandbox has 2 CPUs and 4 GB of memory.
- Bugbearer runs at most 2 browser jobs at a time across the whole service. Cases, reproductions and fix checks share them.

Source: https://docs.bugbearer.com/runners/index.mdx