Nebari Apps Pack

The Nebari Apps Pack lets users launch, manage, and observe static web apps (HTML/CSS/JS, served by nginx) on a Nebari Kubernetes cluster — behind Keycloak SSO, with no Kubernetes knowledge required. Python services are handled by the separate python-capability-pack.

Everything converges on one declarative resource: the App custom resource (apps.nebari.dev/v1alpha1). Whether an app is launched from the web UI, the REST API, or kubectl apply, the apps-operator reconciles the same contract into a Deployment, a Service, and a NebariApp — and the nebari-operator turns that into routing, TLS, authentication, and a landing-page tile.

Each app gets its own URL under the cluster’s apps domain:

https://<subdomain>.apps.<cluster-domain>

What ships today

apps-operator — reconciles App resources. Static apps from inline files, a git repository, or a PVC.
apps-api — REST API for CRUD, status, logs, events, analytics, and direct zip/.html upload.
apps-ui — a dashboard built on the Nebari design system: analytics, an apps table, per-app detail pages with live logs and events, and a launch form with drag-and-drop upload.
apps-mcp — an MCP server at /mcp so coding agents launch and manage apps with natural language (Keycloak device-flow auth).
new-nebari-app skill — teaches Claude Code to scaffold apps in the expected layout with a nebari-app.yaml manifest, so “generate, then launch it” is one flow.

In this guide

Getting started — install the chart on a Nebari cluster and launch your first app
Launching apps — every way to launch: UI, upload, API, and plain CRs
MCP server — connect a coding agent and launch apps with natural language
Scaffolding skill — generate launch-ready apps with Claude Code
Local development — run the whole stack on your laptop with kind

Reference pages

App CRD reference — complete field-by-field reference for the App custom resource
REST API — the apps-api endpoints
Architecture & auth — how the pieces fit together and how authentication works

Looking for Python apps?

Python app support was removed from this pack in favor of python-capability-pack, which owns Python services. See docs/PLAN.md in the repository for the full roadmap.