For hardware makers
We are looking for KVM, camera, serial-console, and input-device makers willing to test how an AI agent can observe and act on a physical machine. The GL.iNet Comet is our working reference. Other devices need their own hardware tests before we claim support.
How we report results
An evaluation unit does not determine support status. We run a device through the same checklist, record its firmware and operating limits, and report what passed and what did not. We share the findings for factual review before publication.
What we need from you
1. Evaluation hardware
An evaluation unit on agreed loan or gift terms. Physical testing lets us document actual behavior instead of inferring support from API documentation.
2. Engineering contact
A technical contact to help answer API questions, review quirks, and provide guidance on firmware update behaviors.
3. API documentation
Public or developer specifications for the relevant channel, such as video, keyboard and mouse, serial text, camera media, or power controls.
What you get
Tested driver
If evaluation succeeds, we plan to build a driver and include it in the future public repository with tests and documented limits.
Public compatibility report
A public report of the checks we can run for the device: observation quality, input accuracy, recovery behavior, and firmware limits where applicable. We will state what was not tested.
Advance review
We will share draft findings with your engineering team before publication so factual errors and firmware details can be checked.
A path into agent workflows
A verified driver could make your hardware usable in agent workflows that need access before an operating system or remote software service is available.
Get in touch
Outreach and technical partnerships are managed by Cisco Caceres. If you would like to arrange an evaluation unit or review driver integration, please use the Bootscry contact page:
Maintainer contact
Cisco Caceres
Bootscry enquiries: Contact Bootscry
Applied AI background: ciscocaceres.com/ai
Agent coordination reference: echo.cc (advisory locks, verified tasks, and agent messaging)
Hardware evaluation: Evaluation units and driver integration requests welcome.