Skip to content

v0.2.5

Released: 20 November 2026

Overview

Tycoon CLI v0.2.5 expands MotherDuck support, introduces a more complete data observability experience, and improves the reliability and security of everyday data workflows.

What's new

MotherDuck integration

  • Ingestion commands can now load into a MotherDuck raw database using md: or motherduck: targets.
  • Semantic scaffolding and post-transform OSI scaffolding can read from a MotherDuck warehouse.
  • A dedicated database-tycoon[motherduck] extra provides the additional dependency required for MotherDuck raw loads.

Data observability

  • A project-level metadata database records dlt ingestion and dbt transformation activity.
  • tycoon data history and tycoon data history show provide terminal access to run history and details.
  • tycoon doctor reports observability capture health.
  • Automatically generated Rill dashboards provide visibility into dlt loads, dbt runs, model outcomes, and test results.
  • tycoon data status now includes run counts for data sources.

Improved data inspection

  • tycoon data layers provides an overview of sources, staging, intermediate, and marts.
  • tycoon data sources explore displays source schemas, sample rows, and run history.
  • tycoon fire and tycoon firehouse surface failing tests in the terminal.
  • tycoon data query --format supports CSV, JSON, and Markdown output.
  • tycoon data schema supports filtering by database and schema.

Reliability and security

  • dlt state is isolated per project to prevent projects with similarly named sources from sharing incremental state.
  • MotherDuck commands check for authentication before connecting, avoiding unexpected browser sign-in flows in scripts and CI.
  • MotherDuck tokens are redacted from CLI output and logs.
  • Local secret files are excluded from version control by default.
  • Dependencies and the PyPI publishing action have been updated to address security advisories.
  • The publishing workflow now verifies that a release tag points to a commit on main.

Important notes

  • MotherDuck raw loads require the destination database to exist, a configured MotherDuck token, and the database-tycoon[motherduck] extra.
  • Switching an existing project from a local raw database to MotherDuck requires updating the dbt profile as well as the Tycoon configuration. dlt incremental cursors start over in the new destination; existing local rows are not copied automatically.
  • Run-history capture for MotherDuck raw loads is not yet supported. Local raw database workflows retain their existing capture behaviour.
  • The minimum supported uv version for contributors is now 0.12.22. This does not affect installing Tycoon from PyPI.