What is stardelt

stardelt is a Kubernetes-native, fully open-source data platform — an opinionated collection of upstream OSS services shipped together as a lakehouse, batch, streaming, notebook, and BI environment that runs on any cloud or on-prem cluster, under your own sovereignty, without vendor lock-in.

What makes stardelt different

• Total sovereignty. stardelt runs end-to-end inside your perimeter. Nothing leaves your cluster unless you tell it to. EU companies escape US CLOUD Act exposure; defense and public-sector workloads can run air-gapped. Reversible by design — your tables stay in open Iceberg format on your storage.
• No vendor handcuffs. 100 % OSI-permissive (Apache 2.0 / MIT / BSD). Zero BSL, SSPL, ELv2, or AGPL components.
• One CRD, not 47 YAMLs. The stardelt Operator reconciles a small set of top-level CRDs into the per-service operators below. That's the gap left by the previous generation of Kubernetes data stacks.
• Modern stack, no Hadoop heritage. Apache Iceberg via Lakekeeper, Trino for interactive SQL, Spark Connect for distributed compute, JupyterHub for notebooks, Superset for BI, Apache Kafka (KRaft) for streaming, SeaweedFS for object storage, Airflow for orchestration — composed by the stardelt Operator and surfaced through stardelt Nova.

Where to go next

• Try it on your laptop. Run stardelt locally on kind.
• Understand the system first. Architecture overview or the full design spec.
• See what's planned. Roadmap.

Pre-alpha

stardelt is in an MVP / vibecoding phase. APIs, manifests, and chart versions change without notice. Not ready for production.