MotherDuck

MotherDuck is a serverless data platform built on DuckDB. Kestra runs the pipelines that load, transform and query it, with schedules, retries and secrets built in.

  • Technology partner
  • United States
  • North America

About

Kestra's DuckDB plugin connects to MotherDuck through an md: connection string and a MotherDuck token, so every DuckDB task and blueprint can run against a shared cloud database. Anna Geller covered the ETL patterns in a guest post on the MotherDuck blog, and DuckDB vs. MotherDuck explains when to move a local workload to the cloud.

MotherDuck + Kestra: cloud DuckDB, orchestrated

  • Same SQL, laptop to cloud

    Prototype a flow on local in-memory DuckDB, then point the same Query task at MotherDuck by changing the JDBC URL. No SQL rewrite.

  • Tokens stay in secrets

    The MotherDuck service token is stored as a Kestra secret and referenced in the JDBC URL, so it never appears in flow code.

  • Results that outlive the run

    In-memory DuckDB disappears when the execution ends. Writing to a MotherDuck database keeps tables between runs and shares them with the team.

  • Scheduled and retried loads

    Cron schedules, retries and backfills run MotherDuck queries on time, and failed runs replay from the task that broke.