Skip to main content

Admiral · Documentation

The operating system
for physical AI

Build, enrol, deploy and operate fleets of robots and edge devices. Everything you need to run Admiral, from the first boot to ten thousand devices.

Get started API reference

01 How it works

No site visits, platform team or bricked robots.

  1. (1)

    Build

    Your Dockerfile becomes an immutable, signed image for your hardware: x86, ARM or RISC-V.

    Configurations →
  2. (2)

    Enrol

    Devices set themselves up on first boot and pair to your organisation, with no one on site.

    Install and pair →
  3. (3)

    Deploy

    Staged rollouts install to a spare partition. The health check passes, or the device rolls back.

    Rollouts →
  4. (4)

    Operate

    Live health, logs and outbound-only remote access, with security updates for the life of the device.

    Device management →

02 Explore the docs

Find your way around.

01

Getting started

Install Admiral OS, pair your first device and deploy a workload.

02

Core concepts

Configurations, fleets, desired state, networking and the platform boundary.

03

Deploying and updating

Rollouts, maintenance windows and A/B OS updates.

04

Hardware and robotics

Device access, ROS 2, GPUs and NPUs, and hardware guides.

05

Operating devices

Monitor, diagnose and repair devices in the field.

06

Security and access

Device security, audit, users, groups and permissions.

07

Automation

The CLI, service accounts and webhooks.

08

API reference

Every endpoint the dashboard uses, for your own tooling.