Skip to main content

API reference

The H2O Workflows Python SDK (h2o-workflows) provides a client library for managing workflows, runs, and related resources in H2O AI Cloud. You can use the SDK to create, update, activate, and trigger workflows programmatically.

Installation​

pip install h2o-workflows

Requires Python 3.8 or later.

What the SDK provides​

Calling login() returns a Clients object that bundles typed clients for each Workflows resource:

  • .workflow (WorkflowClient) — Create, update, activate, deactivate, delete, and validate workflows.
  • .workflow_revision (WorkflowRevisionClient) — List and get workflow revisions.
  • .run (RunClient) — List, get, cancel, delete, and retry runs.
  • .run_attempt (RunAttemptClient) — List and get run attempts.
  • .job (JobClient) — List and get jobs.
  • .step (StepClient) — List and get steps.
  • .runner (RunnerClient) — List and get runners.

Quick start​

import h2o_workflows

# Connect to H2O AI Cloud (uses H2O CLI config by default)
clients = h2o_workflows.login()

# List workflows in a workspace
page = clients.workflow.list_workflows(parent="workspaces/my-workspace")
for workflow in page.workflows:
    print(workflow.name)

Next steps​

See Python SDK Reference for complete method documentation.

Feedback