ARK - Agentic Runtime for Kubernetes

Ark is a set of open-source tools that extend Kubernetes with custom resource definitions (CRDs)  and services for building and running agentic workloads. This allows teams to extend their existing Kubernetes infrastructure to run AI agents, in a similar way to how Argo  extends Kubernetes to run workflows.

Ark provides resources for:

• AI model configurations.
• Autonomous agents.
• Multi-agent teams.
• Tool integrations.
• Memory persistence.

Ark emphasizes a declarative approach to building agentic solutions, separating the definition of how agents should behave from the underlying implementation of how they are run. This encourages teams to build portable solutions that can be moved across different stacks as the agentic space evolves.

Find out more about the core concepts of Ark.

Get started

Follow the Quickstart guide to install ARK locally.

Credits

The initial design and implementation of Ark was led by Roman Galeev , Dave Kerr , and Chris Madden .