Complete this 5-minute setup to start automating with AI agents.

​
Before you start

You’ll need:
  • Access to your Kubiya organization
  • Admin permissions (to connect integrations)
  • Your infrastructure tools ready (AWS, GitHub, Kubernetes, etc.)

​
Step 1: Log in and explore

  1. Go to https://compose.kubiya.ai
  2. Sign in with your organization credentials
  3. Select your organization if you have multiple
You’ll see the main interface with:
  • Left sidebar - Navigation to Tasks, Workflows, Context Graph, etc.
  • Main area - Task creation and activity feed
  • Quick actions - Create Task, Browse Workflows, etc.
Task Creation Interface Main Composer interface with task templates and suggestions

​
Step 2: Try your first task

  1. Click “Create Task” in the main interface
  2. Type this example request:
“Show me what Kubernetes clusters I have access to and list the pods in the default namespace”
  1. Press Enter and watch the AI agent:
    • Connect to your Kubernetes tools
    • Execute the commands safely
    • Return live results with explanations
The agent will guide you through any missing permissions or connection issues.

​
Step 3: Connect your tools

To unlock Composer’s full power, connect your infrastructure tools:
  1. Go to the Resources tab in the left sidebar
  2. Click “New Stack” to add integrations
  3. Choose your platform:
    • AWS - Connect your cloud resources
    • GitHub - Integrate repositories and CI/CD
    • Kubernetes - Manage clusters and workloads
    • Slack - Enable notifications and collaboration
Integrations Overview Browse and configure integrations with popular services Pro tip: Start with one tool you use daily. You can add more later.

​
Step 4: Create a custom agent

Enhance your automation with specialized agents:
  1. Click the agent dropdown in the task interface
  2. Select “AI Engine Configuration”
  3. Choose “Import Agent Preset” and then “Generate with AI”
  4. Describe your needs: “agent that can interact with kubernetes clusters”
  5. Review the generated configuration and click “Accept & Use Agent”
AI Agent Generation Process Let AI create specialized agents based on your requirements Your custom agent will appear in the interface, ready for specialized tasks.

​
Step 5: Build your first workflow

Turn your successful task into a reusable workflow:
  1. Go to Workflows → Create New Workflow
  2. Drag components from the left panel to the canvas
  3. Connect them to define your automation flow
  4. Test it with the “Execute Workflow” button
Visual Workflow Editor Drag-and-drop workflow builder - no coding required Example workflow: “Deploy on GitHub push”
  • Trigger: GitHub webhook
  • Action 1: Build Docker image
  • Action 2: Deploy to Kubernetes
  • Action 3: Send Slack notification
Complex Workflow Execution Watch your workflow execute in real-time with detailed progress tracking

​
Step 6: Explore your infrastructure

Use the Context Graph to understand your connected systems:
  1. Click “Context Browser” in the left sidebar
  2. Watch Composer discover your infrastructure automatically
  3. Explore the network view to see how everything connects
Context Network Interactive map showing relationships between all your connected resources This helps you:
  • Understand dependencies before making changes
  • Find related resources when building workflows
  • Get AI suggestions based on your actual infrastructure

​
You’re ready! 🎉

That’s it! You now know how to:
  • ✅ Chat with AI agents to execute tasks
  • ✅ Connect your infrastructure tools
  • ✅ Create custom specialized agents
  • ✅ Build visual workflows
  • ✅ Explore your system relationships

​
What’s next?

Start automating: Try these common use cases:
  • “Deploy my latest commit to staging”
  • “Scale up my production pods to 5 replicas”
  • “Create a PR for the bug fix in feature-branch”
  • “Show me which services are using the most CPU”
Explore more:

​
Need help?

Ready to transform how your team handles infrastructure? Start with a simple task and see the magic happen.