Legacy Python CLI command reference for AI Runtime

important

This documentation has been retired and might not be updated.

The Python-based air CLI, installed with the databricks-air package, is now deprecated and is no longer actively maintained.

Use the Databricks CLI for new workloads. See Use the Databricks CLI with AI Runtime.

The air CLI provides five subcommands for managing training workloads: air run, air get run, air list runs, air logs, and air cancel. The same reference is always available from the CLI itself:

Bash

air --help              # All commands
air <command> --help    # Flags for a specific command
air -h config           # YAML config reference
air -h config.compute   # Per-field help for a YAML section

The CLI help reflects the exact version you have installed, so it is the source of truth if it differs from the table below.

Commands

Command	Purpose
air run	Submit a workload defined by a YAML file. Supports --file, --watch, --dry-run, --override, --idempotency-key, --email, and -p to select an authentication profile.
air get run	Show metadata, status, and configuration for a specific run.
air list runs	List recent runs. Use --limit to bound the result count and --active to show only running workloads.
air logs	Stream or download logs for a run. Use --node to target a specific node and --download-to to write logs to a directory.
air cancel	Cancel a running workload.
air register image	Register and cache a custom Docker image for use in workloads. See Use custom Docker images with the legacy Python CLI.

Command	Purpose
air run	Submit a workload defined by a YAML file. Supports --file, --watch, --dry-run, --override, --idempotency-key, --email, and -p to select an authentication profile.
air get run	Show metadata, status, and configuration for a specific run.
air list runs	List recent runs. Use --limit to bound the result count and --active to show only running workloads.
air logs	Stream or download logs for a run. Use --node to target a specific node and --download-to to write logs to a directory.
air cancel	Cancel a running workload.
air register image	Register and cache a custom Docker image for use in workloads. See Use custom Docker images with the legacy Python CLI.

Global flags

Flag	Purpose
--version	Print the installed CLI version.
-p, --profile	Use the named Databricks CLI authentication profile instead of the default.
-h, --help	Show help. When followed by a config path (-h config.compute), show YAML field help.

Flag	Purpose
--version	Print the installed CLI version.
-p, --profile	Use the named Databricks CLI authentication profile instead of the default.
-h, --help	Show help. When followed by a config path (-h config.compute), show YAML field help.

See also

• Legacy Python CLI quickstart for AI Runtime
• Workload YAML reference for the legacy Python CLI
• Track runs with the legacy Python CLI
• Use custom Docker images with the legacy Python CLI