ZenML
the six Ws · specification
Maintained by ZenML GmbH.
An MLOps framework for portable, reproducible ML and LLM pipelines across orchestrators and clouds.
A Python library with a self-hostable server, plus an optional managed cloud.
Reach for it when you want tooling-agnostic pipelines from local to production.
Decouples pipeline logic from infrastructure so stacks swap without rewrites.
Integrates with Kubeflow, Airflow, MLflow, and SageMaker.
Maintained by ZenML GmbH; open-source (Apache-2.0) core, self-hostable, with a paid cloud (open-core). Self-hosted keeps data on your infrastructure.