Skip to content

Quick Start

Get from zero to a running Kubernetes workflow in under 5 minutes.

1. Install the MCP server

bash
brew install tiny-systems/tap/mcp-server

No accounts, no sign-ups, no SaaS dependencies.

Alternative install methods

Go install: go install github.com/tiny-systems/mcp-server@latest

Binary download: GitHub Releases

2. Install CRDs and a module on your cluster

You need a running Kubernetes cluster with kubectl configured. Then install the Tiny Systems CRDs and at least one module:

bash
# Add the Helm repo
helm repo add tinysystems https://tiny-systems.github.io/module/

# Install CRDs (required once per namespace)
helm install tinysystems-crd tinysystems/tinysystems-crd \
  --namespace tinysystems --create-namespace

# Install the common module (ticker, cron, router, debug, etc.)
helm install common-module tinysystems/tinysystems-operator \
  --namespace tinysystems \
  --set controllerManager.manager.image.repository=europe-west2-docker.pkg.dev/tinyplatform/tinysystems/common-module-v0 \
  --set controllerManager.manager.image.tag=0.1.236 \
  --set fullnameOverride=common-module

Finding modules

Use search_modules in the MCP server to browse the public module catalog. Each module's get_module_info response includes the exact Helm command to install it.

3. Connect to Claude Code

bash
claude mcp add tinysystems mcp-server serve --namespace tinysystems

4. Build your first flow

Start Claude Code and describe what you want:

bash
claude

Then type:

Show me what modules are installed, then build a flow that ticks every 5 seconds and logs the output to a debug node.

The AI will:

  1. Call list_modules to see what's available
  2. Call get_component_info for ticker and debug
  3. Call build_flow to create the TinyNode CRDs
  4. Return the generated node IDs

Verify in your cluster:

bash
kubectl get tinynodes -n tinysystems
kubectl get tinyflows -n tinysystems

5. Explore solutions

Before building from scratch, check if someone already built what you need:

Search for Kubernetes pod monitoring solutions

The AI calls search_solutions, finds matches, and can adapt an existing solution to your needs.

What's next

Build Kubernetes workflows with a prompt