Solo.io, a cloud-native application networking company, has released kagent, an open source framework designed to accelerate workflows in Kubernetes by enabling AI-driven automation. The framework provides tools, AI agents, and a declarative API to help DevOps and platform engineers build, run, and integrate AI agents seamlessly into cloud native environments.

Kagent introduces a first-of-its-kind agentic AI approach to Kubernetes, allowing teams to automate tasks such as configuration, troubleshooting, observability, and network security. It integrates with existing cloud native tools through the Model Context Protocol (MCP), simplifying large-scale deployments without requiring deep expertise in each tool.

“With kagent, teams can offload undifferentiated heavy lifting to agentic infrastructure and focus on higher-value tasks,” said Keith Babo, Chief Product Officer at Solo.io. Analysts see this as a significant advancement in AI-driven cloud operations.

The framework consists of three key layers:

• Tools: Predefined functions for AI agents, including monitoring, deployment management, and security controls.
• Agents: Autonomous systems that execute complex tasks like canary deployments, zero-trust security enforcement, and debugging.
• Declarative framework: A simple API and controller for managing agents via UI, CLI, and configuration files.

Built with extensibility in mind, kagent launches with integrations for Argo, Helm, Istio, Kubernetes, and Prometheus. The framework leverages Microsoft’s AutoGen and can be expanded with any MCP-compatible tool. As an open source project, kagent aims to drive adoption across the cloud native ecosystem, empowering teams to build, share, and run AI-driven solutions efficiently.