Manage Databricks without the UI
LazyDatabricks is a keyboard-first TUI for clusters, jobs, pipelines, SQL warehouses, and logs. Fast, focused, and safe-by-default with Armed Mode for destructive actions.
| Name | State | Workers | Runtime | Idle | |
|---|---|---|---|---|---|
| my-dev-cluster | TERMINATED | 0 | -- | -- | |
| shared-analytics | RUNNING | 4 | 4h 29m | 5m | |
| job-run-82a4f-xyz | PENDING | 2 | -- | -- | |
| ml-training-prod | RUNNING | 8 | 1h 52m | 12m |
Terminated (inactive for 10 minutes)
Everything you need
A keyboard-first way to run Databricks ops. Fast, deliberate, and built for confident execution.
Cluster Management
Start, terminate, and restart clusters. Monitor state, workers, and runtime at a glance.
Job Monitoring
View job runs, trigger new runs, and cancel running jobs. Three-pane hierarchy for fast navigation.
Pipeline Management
Start, stop, and monitor Delta Live Tables pipelines. Full refresh support.
SQL Warehouses
Monitor SQL warehouse state. Start and stop warehouses on demand.
Log Viewer
View logs with search and filtering. Jump between matches, filter by level.
Armed Mode
Safe-by-default. Destructive actions require explicit arming with a 30-second timeout.
Get started in seconds
Install with pip and launch.
pip install lazydatabricks
lazydatabricksRequires Python 3.10+ and a Databricks workspace
Keyboard-first design
Vim-inspired keybindings for fast, efficient navigation.
About
The Databricks UI can be sluggish when you're bouncing between clusters, jobs, pipelines, and logs. LazyDatabricks is a keyboard-first way to run Databricks ops: fast, deliberate, and safe-by-default with Armed Mode for anything destructive.
Maintainer: Fahad Baig
Related: Cognition Kernel Research
Want to collaborate, integrate with MCP tooling, or share feedback? Open an issue or reach out.