# Device

Device: A machine you've registered, your laptop, a spare workstation, or a cloud instance, that CodeHerder can actually dispatch task work to.

Source: https://codeherder.com/glossary/device/

Glossary · The machinery

A machine you've registered, your laptop, a spare workstation, or a cloud instance, that CodeHerder can actually dispatch task work to.

## What it is

A device is any machine running the CodeHerder device software, registered once so the platform knows it exists and can send it work. It might be a developer's own laptop, a workstation sitting idle in an office, or a cloud instance spun up for exactly this purpose. Whatever it is, once it's registered, tasks can be routed to it, and it stays registered until someone deliberately retires it. A subscription lapsing doesn't remove it.

## Why it matters to you

Devices are where the actual compute, and the actual bill for that compute, lives. Registering the machine you already own instead of paying for a vendor's cloud VM per task is often the difference between an unpredictable line item and a cost you already control. It also means capacity grows the way your team already grows: add a machine and register it, and it's available to pick up work that same afternoon.

## How it works in CodeHerder

CodeHerder tracks each device's status: online or not, its load, its concurrency. It lets you pause or retire one without losing track of the work it was doing. Connecting a machine, either by running the CLI yourself or launching one in your own AWS account, is covered in the adding a device doc. The managing your devices doc covers day-to-day device management: checking capacity and pausing a busy machine.

- [Managing your devices →](https://codeherder.com/docs/devices/)
- [Adding a device →](https://codeherder.com/docs/adding-a-device/)

## What people get wrong

A device isn't required to be dedicated hardware sitting somewhere just for this, even though that's the first picture that comes to mind. It can be the same laptop a developer already uses for everything else, registered so it's also available to pick up task work when it's free. The other assumption worth correcting is that more devices automatically means more work gets done; a device only helps if there's an agent capability registered on it that matches what your backlog actually needs, which is a separate thing to check than simply whether the machine is online.

Keep going

## Related terms

[Agent isolation Running each task's agent in its own copy of the repository, so two tasks never collide by editing the same files at the same time.](https://codeherder.com/glossary/agent-isolation/) [Agent session One running instance of a harness working a task on a device, the actual process you can watch in real time.](https://codeherder.com/glossary/agent-session/) [Capability routing Matching a task to an agent that's actually tagged for that kind of work, instead of the first agent that happens to be free.](https://codeherder.com/glossary/capability-routing/)

[← See every term](https://codeherder.com/glossary)