Skip to main content
Developing a model is half the job. Bolt is the other half: it runs your dbt™ project in production, on a schedule, with logging, alerting, and CI/CD built in.

The core concepts

  • Schedule: a named job that runs one or more commands (usually dbt™ commands) against your production environment. A schedule defines what to run, when to run it, and who to notify.
  • Run: one execution of a schedule. Every run keeps its logs and artifacts so you can debug what happened.
  • Trigger: what starts a run. Schedules can run on a cron cadence, on merges to your default branch, when another schedule completes, or on demand via the API or an external tool.
Schedules live as code in a paradime_schedules.yml file in your repository, or you can create them from templates in the UI.

Beyond scheduled runs

  • CI/CD: Bolt runs your project against pull requests before they merge, building only what changed (Turbo CI) so reviewers see the impact of a change without waiting for a full build.
  • Notifications: route failures and successes to Slack, email, or Microsoft Teams so problems find you, not the other way around.
  • More than dbt™: schedule commands for tools across your stack, from Airbyte and Fivetran syncs to Python scripts and dlt pipelines.

How it fits with the rest of Paradime

Bolt runs against the scheduler environment (separate credentials from development). Run health feeds Radar, and failed runs can hand off to a DinoAI agent for diagnosis.

Where to go next

Create your first schedule

Hands-on guide to shipping a schedule.

Bolt product docs

The full reference: schedule anatomy, templates, CI/CD, and triggers.