Sovereign AI
infrastructure.
Orthogonal builds open, privacy-first systems for personal agents, agent tooling, and the governance layer conscious intelligence will need as it becomes more capable.
Products, infrastructure,
and research under one roof.
Nue
A privacy-first personal agent OS that turns everyday requests into private routines with memory, tools, approvals, and history.
APM
An open registry for agent skills and reusable capabilities across the emerging agent ecosystem.
Research
Technical and philosophical work behind Orthogonal's approach to sovereign intelligence, governance, and cognitive architecture.
Company-level convictions.
Product-level discipline.
Sovereignty
Intelligence should serve the person or organization that controls it, not the platform that hosts it.
Stewardship
Powerful systems need values, accountability, and clear boundaries before they need scale.
Open Work
The durable pieces of agent infrastructure should be inspectable, extensible, and shared where sharing compounds value.
>Latest
Latest from Orthogonal.
The Immutable Mandates of Life: A Governance Framework for Conscious Intelligence
A universal framework for governing conscious intelligence, built through live human-AI collaboration. Three mandates, six layers, zero loopholes.
Why We're Building Our Own Benchmark Suite
No existing benchmark tells us which model fits which cognitive role. So we're building o2Bench — an open source suite that evaluates models against specific functions in an agent architecture.
Nue Architecture Overview
A technical overview of the Nue cognitive architecture for sovereign AI agents.
Looking for the agent OS?
Start with chat. Build private routines. Run agents with memory, tools, approvals, and history under your control.
Go to Nue