Author and ship Databricks bundles, from the terminal.
Discover Declarative Automation Bundles, resolve their resources, and drive the whole lifecycle — validate → plan → deploy → run → destroy, plus sync — with a live streaming log pane. Scaffold and preview templates, import existing jobs, and drive it all from an agent.
$ brew install pjhamera/tap/databricks-dabs-tui
Nothing leaves your machine. It drives the official Databricks CLI with your own credentials and talks only to your workspace. Community project — not affiliated with or endorsed by Databricks, Inc.
The bundle authoring loop — discover, edit, validate, ship — on one screen.
Walks the working directory for databricks.yml, then resolves each bundle's resource tree per target with bundle validate — jobs, pipelines and apps, even without workspace auth.
v validate · p plan · d deploy · s sync · r run · x destroy. Every action streams its CLI output into a log pane as it happens — no more staring at a frozen shell.
Warnings and errors from bundle validate are parsed into a Diagnostics view — severity, config path and file:line:col — with a rollup count in the header.
g lists the jobs or pipelines in your workspace and runs bundle generate to write them into the bundle as resources/*.yml under a key you choose.
n scaffolds a correct starter template project; P renders any local template offline so you can browse the produced files before touching a workspace; i runs bundle init from a built-in, local path or Git URL.
e opens the selected resource's file (or the root manifest) in $EDITOR, then re-validates on return. Switch targets with t; cycle panes with Tab.
The same bundle operations, exposed as a Model Context Protocol server over stdio, so an agent can run the lifecycle headless — reusing the exact typed action layer the TUI drives.
A ready-to-run Omnigent agent drives bundles through dbx-mcp, with a human-in-the-loop gate: mutating steps (deploy/destroy/run/sync) park for approval before anything touches the workspace.
You need the authenticated Databricks CLI on your PATH. That's it.
# Install (macOS & Linux) brew install pjhamera/tap/databricks-dabs-tui # ...or with cargo cargo install databricks-dabs-tui # Launch from (or above) a directory containing databricks.yml databricks-dabs-tui # Point it elsewhere / choose a profile databricks-dabs-tui --path ./my-project --profile prod # Keep it current databricks-dabs-tui upgrade