stardelt
A self-hosted data platform for your own Kubernetes: Iceberg tables, SQL, pipelines and dashboards behind one sign-in — no hyperscaler needed.
Project intro
stardelt is a self-hosted data platform that puts Iceberg tables, SQL, pipelines, notebooks, streams and dashboards behind one sign-in, one interface and one API, running inside your own Kubernetes cluster instead of a hyperscaler's cloud.
It is built from open-source pieces the author picked himself: Iceberg tables on your own S3 storage, Trino for SQL, Spark Connect for jobs, Airflow for schedules, Kafka streams, plus OIDC login and OpenFGA, all tied together by a Rust control plane called Nova. The site walks through a normal day with a mobility team: taxi trips land as a file at 7:00 and by 9:00 the team has a modelled table, a daily schedule, lineage and a dashboard, without leaving the one interface.
The author is a data platform engineer in Germany who spends his days running lakehouse platforms on US hyperscalers for small businesses and DAX companies, and started stardelt in his spare time because he was unhappy with vendor lock-in, high cloud bills, non-EU vendors and the licence fees other open-source tools charge for LDAP. It is open source under Apache 2.0.
Be clear-eyed about the state of it: the site itself calls it pre-alpha and says it changes quickly, and the GitHub organisation describes it as an MVP in a vibecoding phase that runs end to end on a single-node k3s test cluster but is not ready for any production use. The Show HN post went up on 8 November 2026 and the organisation's repos have one star at most, so this is a very early project to watch or try on a spare cluster, not something to bet a pipeline on.
Actions
Tech stack
Project info
- Project category
- Developer Tools
- Project status
- Online
- Collaboration intent
- None