Dagster is a system for building modern data applications. Dagster provides an elegant programming model, UI tools, and programmatic APIs to enable infrastructure engineers, data engineers, and data scientists to collaborate on ETL/ELT, ML pipelines, building data applications, and more.

Elegant Programming Model

Dagster is a set of abstractions for building self-describing, testable, and reliable data applications. It embraces the principles of functional data programming; gradual, optional typing; and testability as a first-class value.

Beautiful Tools

Dagster's development environment, dagit — designed for data engineers, machine learning engineers, data scientists — enables astoundingly productive local development.

Flexible and Incremental

Dagster integrates with your existing tools and infrastructure. Dagster can invoke any computation — whether it be Spark, a Python, a Jupyter notebook, or SQL — and is designed to deploy to any workflow engine, such as Airflow.