Agent Runtime
Agents operate through explicit capabilities and persistent task state.
KINGAI OS Server
Intelligence for infrastructure you control.
Run the KINGAI governed runtime on servers, VPS and private AI nodes without turning a model into root. Policy, approval, tasks, memory, model routing, audit and recovery stay part of the system.
Server Developer Preview · amd64 public ISO0.1.0-dev
Headless by design
The intelligence layer without desktop overhead.
Server focuses on governed agent runtime, model routing, persistent tasks, local memory, system health and recoverable operations.
Policy & Approval
Identity, target, capability and risk stay outside model authority.
Memory & Models
Local memory metadata and model-neutral routing support private AI infrastructure.
Recovery
Installer, A/B update, rollback and recovery paths are part of the lifecycle work.
Developer ready
Built for people who operate systems.
Linux tooling, Git, Node.js, browsers and AI automation are part of the broader KINGAI development direction.
$ kingai status Runtime: D5 Policy: active Audit: enabled $ kingai doctor System diagnostics and bounded repair workflow
CLI surfaces shown here reflect current project direction and may evolve with the runtime.
Recoverable operations
Move forward without losing the last known-good system.
Server engineering treats update, health confirmation, rollback and recovery as one operating lifecycle.
Run AKnown-good active slot
Stage BWrite the inactive slot
Health confirmationConfirm only after successful boot
Rollback / RecoveryReturn to a verified state on failure
KINGAI OS Server
Put governed intelligence where your infrastructure lives.
The current public artifact is a Developer Preview; newer source capabilities may be ahead of the published ISO.