Cognitum
Cognitum One announces ruOS, a governed computer for teams of AI agents
On October 6, Cognitum One announced ruOS: managed Linux workspaces where teams of agents work on infrastructure the enterprise controls, built on MCP and available as a .deb package.
Announced: · Published:
Cognitum One announced ruOS on October 6. It is the execution layer for companies that want AI agents with computers of their own, without handing those computers to someone else's cloud.
According to the announcement, ruOS runs agents inside managed Linux workspaces, locally or on infrastructure the customer controls. Restricted-network and air-gapped setups are supported when the models and dependencies are available locally. It is built on the Model Context Protocol, works with the models an organization chooses, and coordinates many specialized agents through Ruflo, with RuVector keeping task context between steps.
The release compares the idea to the remote desktop: centrally governed computers, now for agents. More information is at ruos.cognitum.one.
Draft history: this summary was prepared for the unpublished October 19 candidate and is intended for first publication with the October 27 candidate; the announcement date is unchanged.
- Reuven Cohen — Founder and Chief AI Officer, Cognitum One
- D. Christopher Keil — Founding CEO, Cognitum One
- Mark B. Templeton — Founding advisor to Cognitum One; former CEO, Citrix Systems
Read the original on Business Wire