Kubernetes-Native Agent Runtime

Run any agent on your Kubernetes platform

kagent extends Kubernetes with the identity, orchestration, and governance layer required for production AI agents.

Try kagent
See the Architecture

The Reality

Your Teams Are Already Building Agents

LangChain in engineering. N8N in ops. Claude Desktop on every laptop. CrewAI experiments in data science. Custom Python scripts everywhere.

This is good. The diversity means people are solving real problems with AI.

The question isn't how they build agents. It's where do those agents run in production?

A technical diagram illustrating Cross-Cluster Connectivity in a multi-cloud or multi-region service mesh. Two Kubernetes clusters, Cluster 1 (US-East-1) and Cluster 2 (EU-West-1), are connected via a Global Service DNS that maps to the address orders.mesh.internal. The diagram shows: Cluster 1: Contains "orders" pods that communicate with the DNS. Cluster 2: Contains "orders" pods that are also part of the same global service. Encryption: Large lock icons on the connection lines represent mTLS encryption and secure communication between clusters. Traffic Flow: A "Client Pod" in a separate namespace is shown accessing the orders.mesh.internal address, which then routes to the available pods in either cluster.

From Cloud-Native to Agent-Native

Your agents need more than a server. They need identity, policy, observability, and lifecycle management — built on the infrastructure you already trust.

LangChain logo
LangChain
Framework
CrewAI logo
CrewAI
Framework
Any Agent logo
Any Agent
Bring your Own
ADK logo
ADK
Framework
Any MCP logo
Any MCP
Connectors
Any Skill logo
Any Skill
Capabilities
A horizontal list of four feature categories—Agent, Observability, Security, and Scale—with their associated capabilities. Agent: Short term memory, Long term memory, Session management, Human-in-the-loop, Streaming, Compression, and Evaluation. Observability: Metrics, Tracing, Logging, Auditability, Explainability, OpenTelemetry, and Human-on-the-loop. Security: AccessPolicy, mTLS, On Behalf Of, Secure Elicitation, Token Exchange, RBAC, and OIDC/Oauth. Scale: Multi-cluster, Fault tolerance, Agent HA, Agent Mesh, Single Pane of Glass, and Lifecycle.A vertical list of four feature categories—Agent, Observability, Security, and Scale—with their associated capabilities. Agent: Short term memory, Long term memory, Session management, Human-in-the-loop, Streaming, Compression, and Evaluation. Observability: Metrics, Tracing, Logging, Auditability, Explainability, OpenTelemetry, and Human-on-the-loop. Security: AccessPolicy, mTLS, On Behalf Of, Secure Elicitation, Token Exchange, RBAC, and OIDC/Oauth. Scale: Multi-cluster, Fault tolerance, Agent HA, Agent Mesh, Single Pane of Glass, and Lifecycle.KubernetesKubernetes

Open Source & Community-Driven

The Fastest-Growing Agent Runtime in Cloud Native

Backed by contributions from the world's leading technology companies. A CNCF Sandbox project — and just getting started.

A grid of corporate logos against a lavender-gray background, each set within a slightly darker rounded rectangle: Google, Solo.io, Confluent, Microsoft, IBM, Adobe, Red Hat, SAP, Strava, Amazon, Oracle, AmdocsA grid of corporate logos against a lavender-gray background, each set within a slightly darker rounded rectangle: Google, Solo.io, Confluent, Microsoft, IBM, Adobe, Red Hat, SAP, Strava, Amazon, Oracle, Amdocs
11

Months

1
K+

Community

295

Contributors

2
K+

GitHub Stars

Cloud Native Computing Foundation

Solo Enterprise for Kagent

Everything Kubernetes needs to Run Agents, MCP Tools, and Skills

Confidently deploy and manage agentic AI in your enterprise environment.

Agent Identity & Policy

Granular access control with agent-on-behalf-of semantics. Define who can do what, and when. Every agent gets a verifiable identity, and every action is governed by policy.

Get started

Observability & Tracing

Full tracing and visualization of every agent action, tool call, and decision path. Automatic origination and propagation of tracing for end-to-end observability with OpenTelemetry.

Get started
A technical diagram illustrating Agent Mesh Traffic Routing in a Kubernetes environment. A Client Pod sends a request to an Agent Service via the DNS name agent.mesh.internal. The request is intercepted by a Waypoint Proxy (Envoy), which performs Layer 7 (L7) functions such as Evaluation and Guardrails. The traffic is then securely routed using mTLS (indicated by a lock icon) to one of several Agent Pods (Pod A, Pod B, or Pod C) within the cluster.

Human-in-the-Loop

Keep humans in control of critical decisions. Define approval workflows, review checkpoints, and escalation paths. Agents propose, humans approve.

Get started

Multi-Framework

LangChain, CrewAI, Google ADK, A2A, or custom frameworks — run them all with unified governance. One platform, regardless of how agents are built.

Get started
A diagram showing a circular workflow or relationship between three entities, connected by a thin purple line forming a large circle on a dark background: Agent (Top): Represented by a robot head icon. Human (Bottom Left): Represented by a person silhouette icon. Tool (Bottom Right): Represented by a ruler and pencil icon. Each entity is housed within its own smaller purple circular pod.

MCP Tools

Deploy, secure, and sandbox MCP tool servers with consistent patterns. Complete lifecycle management for the tools your agents depend on.

Get started

Prebuilt Agents & Tools

Get started faster with production-ready agents and MCP-enabled tools. From Kubernetes management to observability, our registry has you covered.

Browse Registry
Alt text: A technical diagram illustrating Agent Mesh Traffic Routing. A Client Pod sends a request to an Agent Service using the address agent.mesh.internal. The request is intercepted by a Waypoint Proxy (Envoy), which handles Layer 7 functions including Evaluation and Guardrails. The traffic is then securely routed via mTLS (indicated by a lock icon) to one of three Agent Pods (Pod A, Pod B, or Pod C) within the cluster.

Go Deeper

Explore the docs, join the community, and start building.

Documentation

Install guides, API references, and architecture deep-dives.

Read the Docs

Blog

Tutorials, use cases, and the latest from the Kagent team.

Read the Blog

GitHub

Star the repo, file issues, and contribute to the project.

View on GitHub

Discord

Chat with the community, ask questions, and share what you're building.

Join Discord