Skip to main content
Skyhook generates the GitHub Actions, Kubernetes manifests, and GitOps wiring that most teams build by hand. You bring a cluster and a repo; Skyhook fills in the boring middle so engineers can ship to Kubernetes without running a platform team.

Pick your path

I'm shipping a service

Import a repo, get generated workflows + Kubernetes config, deploy to dev in under ten minutes. Start with Create and Import Services.

I'm setting up the platform

Connect clusters, set up GitOps, configure addons and CI/CD credentials. Start with the Onboarding Guide.

How it works

1

Connect a cluster

Bring your own GKE, EKS, AKS, or any Kubernetes-compatible cluster. Skyhook installs a lightweight read-only connector via Helm — clusters stay in your account.
2

Import or create a service

Skyhook scans your repo, detects language and framework, and generates a Dockerfile, Kubernetes manifests, and CI/CD workflows. You review the pull requests before anything merges.
3

Deploy

Push to main and your service builds and deploys automatically. Promote across environments from the UI, CLI, or a workflow_dispatch run in GitHub Actions.
4

Iterate

Logs, metrics, traces, preview environments, GitOps with ArgoCD — all wired up. Override anything the generators produced; it’s your repo.

Set it up from your terminal

The fastest way through onboarding is the CLI:
# install (macOS — see Onboarding Guide for Linux/Windows)
brew install skyhook-io/tap/skyhook

# walk through cluster connect → GitHub OIDC → ArgoCD bootstrap → first service
skyhook onboard
The same flow is available in the web UI at app.skyhook.io. Both produce the same result.

What’s in here

Core concepts

Services, environments, clusters, GitOps, and the difference between the cluster connector and the AI agent.

CI/CD & deployment

How generated GitHub Actions workflows build, deploy, and promote your services across environments.

GitOps with ArgoCD

Optional declarative deployments — Git as the source of truth, ArgoCD reconciling clusters.

Connecting clusters

GKE, EKS, AKS, or any conformant Kubernetes — including bare-metal, k3s, and on-prem.

Observability

Logs, metrics, traces, and FinOps — opinionated defaults you can swap out.

Skyhook Agent

AI assistant that knows your services, clusters, and code. Ask, debug, and configure in chat.
Stuck? Email support@skyhook.io — a real engineer reads it.