kagent-dev/kmcp

CLI tool and Kubernetes Controller for building, testing and deploying MCP servers

What it solves

It simplifies the transition of Model Context Protocol (MCP) services from isolated prototypes to production-ready deployments. It addresses challenges like ad-hoc scaffolding, transport protocol fragmentation (HTTP, WebSocket, SSE), and the lack of consistent security and observability when scaling agent-to-tool communication in enterprise environments.

How it works

kmcp provides a development platform and control plane that integrates with Kubernetes. It consists of three main components:

  • CLI: Used for scaffolding new projects (supporting FastMCP and MCP Go SDK), managing tools, building container images, and local testing.
  • Controller: A Kubernetes-native manager that uses Custom Resource Definitions (CRDs) to handle the lifecycle of MCP server deployments.
  • Transport Adapter: A layer that fronts the MCP server to route external traffic and support multiple transport protocols without requiring code changes.

Who it’s for

  • AI/ML Engineers who need to package and deploy existing MCP prototypes.
  • DevOps Engineers building MCP infrastructure within Kubernetes clusters.
  • Developers looking for a consistent workflow from local development to production.

Highlights

  • Kubernetes-native: Uses CRDs to manage MCP servers as native objects via kubectl.
  • Multi-transport support: Built-in support for HTTP, WebSocket, and Server-Sent Events (SSE).
  • Rapid scaffolding: Quick project setup for Python (FastMCP) and Go (MCP Go SDK).
  • Integrated secrets management: Leverages Kubernetes secrets for secure configuration.

Related

  • Project
    Neargye/hello_tf_c_api

    A collection of C++ examples and a helper library for using the TensorFlow C API to run machine learning models in native applications.

  • Project
    gordonmurray/firnflow

    A multi-tenant vector and full-text search engine that uses object storage as its primary backend with a tiered RAM/NVMe cache for low-latency retrieval.

  • Project
    openai/tunnel-client

    A secure agent that connects private or localhost MCP servers to OpenAI-hosted products like ChatGPT and Codex without exposing them to the public internet.

  • Project
    paulpierre/markdown-crawler

    A multithreaded web crawler that converts website pages into markdown files to simplify document parsing and chunking for RAG and LLM fine-tuning.