MCP Directory

Cyclops MCP

Let AI agents safely create and update Kubernetes applications via Cyclops Modules and templates.

Unverified
stdio (local)
No auth
Stale
Go
View repo 30 Website

Add to your client

Copy the config for your MCP client and paste it into its config file.

Install / run
GOBIN="$HOME/go/bin" go install github.com/cyclops-ui/mcp-cyclops/cmd/mcp-cyclops@latest

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "cyclops-mcp": {
      "command": "mcp-cyclops",
      "args": [],
      "env": {
        "KUBECONFIG": "/path/to/your/kubeconfig"
      }
    }
  }
}

Step-by-step guides: Add to Claude Desktop · Add to Cursor · Add to Windsurf

Before you start

  • A Kubernetes cluster with Cyclops and its CRDs installed
  • Cyclops v0.20.1 or greater (for UI-based MCP install)
  • A kubeconfig (defaults to in-cluster config or $HOME/.kube/config) for the stdio binary
  • Go toolchain (to install the mcp-cyclops binary via 'go install')

About Cyclops MCP

Cyclops MCP is the official Model Context Protocol server for Cyclops, a developer-friendly Kubernetes platform. It gives AI agents tools to manage applications as Cyclops Modules rather than raw Kubernetes manifests: the agent reads available templates and their JSON schemas, then creates or updates Modules that Cyclops renders into Deployments, Services, Ingresses, and more. This keeps changes high-level and validated, reducing the risk of misconfiguration reaching production. The server can run as a local stdio binary (using your kubeconfig) or be installed into a cluster with SSE transport for shared team access; installation through the Cyclops UI is the recommended path (requires Cyclops v0.20.1+).

Tools & capabilities (7)

create_module

Create a new Module. Call get_template_schema first to validate values for the given template.

get_module

Fetch a Module by name.

list_modules

List all Cyclops Modules.

update_module

Update a Module by name. Call get_template_schema first to validate values for the given template.

get_template_schema

Return the JSON schema for the given template. Must be checked before calling create_module.

get_template_store

Fetch a Template Store by name.

list_template_store

List Template Stores from the cluster.

When to use it

  • Let an AI agent create production-ready Kubernetes applications as Cyclops Modules instead of hand-writing manifests
  • Safely update existing deployments by validating values against a template's JSON schema before applying changes
  • Inspect available templates and template stores to understand what application shapes the cluster supports
  • Abstract Kubernetes complexity away from developers and agents while keeping configurations consistent
  • Share a single in-cluster MCP server (SSE) so an entire team's agents connect to the same Cyclops backend

Security notes

By default the server uses your local .kube/config (or in-cluster config) to connect to your Kubernetes cluster, so it operates with whatever permissions that kubeconfig/context grants. It can create and update workloads in your cluster — scope the credentials and target namespaces appropriately. The SSE deployment exposes the server on a cluster service (port 8000) with no built-in authentication, so restrict network access when exposing it.

Cyclops MCP FAQ

How do I install it?

The recommended way is through the Cyclops UI under Addon > MCP server (requires Cyclops v0.20.1+). You can also install a local stdio binary with 'go install github.com/cyclops-ui/mcp-cyclops/cmd/mcp-cyclops@latest', or deploy it into your cluster with SSE transport via the provided install/mcp-server.yaml manifest.

What transports are supported?

The local binary runs over stdio using your kubeconfig. When installed into a cluster (via the UI or manifest), it runs with SSE transport and is reachable at an http URL such as http://localhost:8000/sse after port-forwarding the cyclops-mcp service.

How does it connect to my cluster?

By default it uses the in-cluster config or your $HOME/.kube/config. You can override this with the KUBECONFIG environment variable and choose a context with CYCLOPS_KUBE_CONTEXT. Namespaces for modules and Helm releases are configurable via CYCLOPS_MODULE_NAMESPACE, CYCLOPS_HELM_RELEASE_NAMESPACE, and CYCLOPS_MODULE_TARGET_NAMESPACE.

Why use Cyclops Modules instead of raw Kubernetes resources?

Modules are high-level resources rendered from validated templates. Having the agent work at this level (and check the template schema first) reduces the chance of uncaught misconfigurations compared to editing every line of underlying Deployments, Services, and Ingresses.

#kubernetes#cyclops#devops#platform-engineering#k8s#deployment

Alternatives to Cyclops MCP

Compare all alternatives →
GitHub MCP Server
DevOps & Cloud
31k

GitHub's official server for repos, issues, PRs, and Actions — local Docker or hosted remote.

Featured
Verified
stdio (local)
API key
Go
7 tools
Updated 15 days agoRepo
AWS API MCP Server
DevOps & Cloud
9.3k

Official AWS Labs server that lets agents call any AWS API via the AWS CLI surface.

Verified
stdio (local)
API key
Python
3 tools
Updated 17 days agoRepo
kubefwd
DevOps & Cloud
4.1k

Bulk Kubernetes port forwarding with an MCP server that lets AI assistants forward services, inspect pods, and monitor traffic.

Unverified
stdio (local)
No auth
Go
28 tools
Updated 9 days agoRepo

Compare Cyclops MCP with:

vs GitHub MCP Servervs AWS API MCP Servervs kubefwdvs Radar