Get started

DaggerML requires Python 3.11 or later. Install it, then create a project with the CLI.

pip install daggerml

mkdir research-demo
cd research-demo
dml init

{"ahead":null,"behind":null,"branch":"main","branches":[],"commit":null,"mode":"attached","num_indexes":0,"upstream":null}

The commands remain in one shell block so dml init runs inside the new research-demo directory. It creates .dml/ there. The next page uses that same project to create and query a live DAG.

Continue with DAGs. Use dml --help for generated command details; the ordered Use course applies administrative commands in their project, runtime, and sharing workflows.