Software Packs
Software Packs are the apps you install on Nebari. Browse the catalog and pick the capabilities your team needs: dashboards, model serving, chat assistants, document analysis, and more. Every pack plugs into Nebari's single sign-on, routing, and TLS automatically, so users reach them through one secure entry point.
Installing a pack
To install any pack:
- Make sure your Nebari cluster is up and running.
- Install the pack. Every pack's README documents the exact steps; pick one of two paths:
- Recommended: add it to your gitops repo as a new ArgoCD Application that references the pack. ArgoCD then deploys it onto your cluster.
- Alternative: install the pack's chart manually with Helm.
- Open your cluster's landing page; the new capability appears there once it finishes installing.
Official packs
Packs maintained by OpenTeams and listed in the central Nebari registry. Each pack declares a maturity level. See How packs mature below for definitions, and the pack dashboard for the current level of each pack.
| Pack | Description | Source |
|---|---|---|
| Chat | A team chat app with customizable AI agents, ready to use behind your Nebari single sign-on. | nebari-chat-pack |
| Data Science | JupyterHub with room to publish interactive apps alongside notebooks, behind your Nebari single sign-on. | nebari-data-science-pack |
| LGTM Observability | Grafana's LGTM stack (Loki, Grafana, Tempo, Mimir) for logs, metrics, traces, and dashboards on your Nebari cluster. | nebari-lgtm-pack |
| LLM Serving | A managed API for running large language models on your cluster, with per-model access control and rate limits. | nebari-llm-serving-pack |
| Longhorn Backup | Hourly snapshots and daily off-site S3 backups for every volume on your cluster's default storage. | nebari-longhorn-backup-pack |
| MLflow | Experiment tracking and a model registry for your ML team, with Nebari single sign-on. | nebari-mlflow-pack |
| Nebi | Shared, reproducible Python and data science environments managed from a single CLI. | nebari-nebi-pack |
| Pi Coding Agent | AI-assisted coding inside your team's JupyterHub notebook sessions. | nebari-pi-coding-agent-pack |
| Provenance Collector | A compliance scanner that proves what's running on your cluster and where each piece came from. | nebari-provenance-collector-pack |
| Ray Serve | Distributed serving for ML models and Python apps, with autoscaling built in. | nebari-rayserve-pack |
| skillsctl | A CLI and registry for discovering, installing, and publishing Claude Code skills on your Nebari cluster. | skillsctl |
| Superset | BI dashboards and ad-hoc data exploration through a web UI, behind your Nebari single sign-on. | nebari-superset-pack |
All packs above are maintained by nebari-dev and are considered official.
How packs mature
A pack's level reflects how stable, tested, and supported it is.
| Level | What you can expect | Intended for |
|---|---|---|
| Experimental | May not install or work; can disappear at any time | Contributors only |
| Alpha | Happy path works on a current dev cluster; limitations documented | Internal demos |
| Beta | Stable enough for a pilot with engineering support; APIs may still shift | Customer pilots |
| GA (v1.0+) | Fully supported, with a documented upgrade and security-fix story | Production |
A pack at any level can also be marked Deprecated, with a sunset date (when it will be retired) and a migration path. Avoid new deployments.
Community packs
A community pack is a Software Pack maintained outside OpenTeams. It follows the same structure as an official pack and can be installed on any Nebari cluster.
Want to share your pack with the Nebari community? See Build your own pack to get started.