Quick Start

Install Skyflo, configure the required access, and run the first workflow.

Install with curl#

One-line install:

curl -fsSL https://skyflo.ai/install.sh | bash

Install with Helm#

1. Add Helm repo

helm repo add skyflo https://charts.skyflo.ai
helm repo update skyflo

2. Create values.yaml

Create a minimal values.yaml with only the required fields:

engine:
  secrets:
    llmModel: "gemini/gemini-2.5-pro" # or openai/gpt-5.3-codex, anthropic/claude-sonnet-4-6, etc.
    geminiApiKey: "AI..." # set only the key for your chosen provider

See all available values:

helm show values skyflo/skyflo

3. Install

helm install skyflo skyflo/skyflo -n skyflo --create-namespace -f values.yaml

The Command Center (skyflo-ui service) is exposed on port 80 in the skyflo namespace. Access it via your cluster's networking setup (LoadBalancer, Ingress, or port-forward). See the Production Deployment guide for details.

What to Try First#

Run these in the Command Center:

• List all pods in the default namespace
• Show me deployments that restarted in the last hour
• What's wrong with the payments pod?
• Rollback the frontend deployment to the previous version

Next Steps#

• Execution Model: how the control loop works
• Helm Values Reference: full configuration
• Kubernetes Operations: what you can do