Agentic Layer is an orchestration and runtime layer for building and operating multi-agent AI systems in enterprise environments.
It provides reusable components and APIs to coordinate agents, manage execution flows, and integrate with existing systems, without locking you into a specific framework or vendor.
As a lightweight, Kubernetes-native platform add-on, it introduces clear execution boundaries, observability, and control mechanisms for running agent-based workloads in production.
To get started with the Agentic Layer, please visit our documentation. Besides others, there you find
- Tutorials
- Installation guides
- Architecture documentation
We are looking for contributors now! Interested? How about:
- Give the Agentic Layer a try and discuss your experience and ideas.
- Check out the upcoming topics in our roadmap.
- Head over to our Kanban board and make your hands dirty yourself.
- Know a nice new gateway for Agents, MCP, or LLM Models? Implement your own operator based on the agentic-layer CRDs!
- Prefer a more efficient agent framework than ADK with Python? Implement a new reference template.
Agentic Layer projects adopt our Code of Conduct.
Contributions from the community are welcome! Please follow our Contributing Guide.