Skip to content

The systems
behind Arcturus.

Arcturus connects machine observations to investigation and maintenance work. Rubian supplies the programmable working environment; Kestowv supplies the systems execution and coordination foundation.

Keep the investigation
with the machine.

Startup inventory and retained project records give an operator a place to begin. Rubian commands inspect the environment; Kestowv pilot interfaces cover service operation and recovery. Arcturus brings that work into a machine-reliability workflow.

Evaluate performance, recovery and operator control.

Performance checks must retain correct output. Recovery checks must establish that the operator can restore and use the service. We assess those requirements together.

Read the company thesis
The product and its foundation
Reliability & managementArcturus

Machine state, operating history and recovery.

The working environmentRubian

Commands, workspaces, jobs and services.

The systems foundationKestowv

Execution, resources and workload evaluation.

Commands and operating controls

Rubian’s manuals connect persistent working environments, programmable controls and machine execution.

Persistent user and agent spaces

Identities, files, projects and running work stay associated with their user space. The agent VMOS provides a machine-facing place for agents to work alongside people: separate contexts on shared computing resources.

Agent workspaces

Commands, libraries and jobs

Readable commands let people and agents direct work. Background jobs preserve an inspectable lifecycle; Forge makes compatible Ruby libraries available from the environment.

Background jobsCapability reference

Security and network access

Accounts, groups and file permissions define access. SSH/SCP, TCP/UDP checks and socket observations connect the environment to the systems around it. Follow the documented command and permission contracts.

Accounts & permissionsNetworking

Execution, service state and recovery

The Kestowv pilot documents service-operation interfaces and repeatable workload evaluation. Correctness, resource use and useful output belong in the same comparison.

Pilot interfacesWorkload evaluation

Core and Project Valux

Core is the human- and agent-facing interface direction. Project Valux is the broader operating-system ambition.

Manuals and research

The manuals document commands, returned data and operating limits. Research pages describe methods and open questions; deployment records identify contributions used in other software.

Start with an operating problem

Find the capability and commands associated with the problem you’re trying to solve, rather than reading an entire architecture introduction first.

Problems We Solve

Inspect methods and results

Read the test conditions and limits. An upstream adoption record, a workload result and a product demonstration are distinct kinds of evidence.

Research & studiesBenchmark records

Bring the engineering question.

Tell us which system you’re responsible for, what’s changing and what a useful result would look like.