Model Context Protocol (MCP) server that lets AI agents create, list and update Cyclops Kubernetes Modules via a set of tools (create_module, get_module, list_modules, update_module, get_template_schema, get_template_store, list_template_store).
https://github.com/cyclops-ui/mcp-cyclopsStop watching your AI assistant fumble with Kubernetes YAML files and risk production misconfigurations. The Cyclops MCP server gives your AI tools the context and guardrails they need to manage Kubernetes applications properly.
You know the drill: ask your AI assistant to help deploy something to Kubernetes, and it confidently generates YAML that looks right but breaks production. Maybe it's the wrong namespace, missing labels, or incompatible resource limits. Raw Kubernetes manifests give AI agents too much rope to hang themselves with.
Cyclops MCP flips this dynamic. Instead of generating raw YAML, your AI assistant works with validated, high-level Cyclops Modules. Before creating or updating anything, it must fetch the template schema and validate inputs against it. No more "looks about right" deployments.
Here's what happens:
get_template_schema for your application templatecreate_module or update_module with validated dataDeploy a new microservice:
"Deploy the user-auth service using our standard microservice template with 3 replicas, 512Mi memory limit, and the staging configuration"
Your AI assistant fetches the microservice template schema, validates the configuration, and creates a properly configured module.
Scale applications intelligently:
"Check our current modules and scale up anything using more than 80% CPU"
The assistant lists modules, checks current resource usage, and updates configurations based on actual template constraints.
Environment-specific deployments:
"Deploy the analytics dashboard to staging with debug logging enabled"
Template schemas can include environment-specific fields, ensuring deployments get the right configuration for each environment.
Cursor IDE:
{
"mcpServers": {
"cyclops-kubernetes": {
"url": "http://localhost:8000/sse"
}
}
}
Binary installation:
{
"mcpServers": {
"mcp-cyclops": {
"command": "mcp-cyclops",
"env": {
"KUBECONFIG": "/path/to/your/kubeconfig"
}
}
}
}
The installation is straightforward - port-forward the service to localhost:8000 and add the MCP server to your AI assistant's configuration. No complex setup or authentication gymnastics.
Fewer late-night production incidents: Template validation catches configuration errors before they hit your cluster. Your AI assistant can't deploy something that violates your organization's standards.
Faster iteration cycles: Instead of reviewing every line of generated YAML, you review high-level module configurations that map to your team's established patterns.
Consistent deployments: Whether you're deploying manually or asking your AI assistant to handle it, everything goes through the same validated templates and constraints.
Less context switching: Your AI assistant understands your existing applications through list_modules and can make informed decisions about scaling, updating, or creating new services.
The Cyclops MCP server sits at 26 stars on GitHub, which means you're getting in early on a tool that solves a real problem without the usual enterprise complexity. Install it, configure your AI assistant, and start deploying Kubernetes applications with confidence instead of crossed fingers.