CLI Installation

Skill: databricks-config

What You Can Build

The Databricks CLI is the prerequisite for bundle deployments, OAuth authentication, workspace administration from the terminal, and most AI Dev Kit workflows. You’ll install it on your platform, verify it meets the minimum version, and handle sandboxed environments where Homebrew isn’t available.

In Action

“Install the Databricks CLI on macOS using Homebrew and verify the version meets the minimum for bundle support.”

# Install via Homebrew
brew tap databricks/tap
brew install databricks

# Verify version (must be >= 0.292.0)
databricks --version

Key decisions:

• Homebrew on macOS is the recommended install method — it handles PATH configuration and makes upgrades a single brew upgrade databricks command
• v0.292.0 or later is the minimum version for bundle features, OAuth login, and the full CLI command surface
• After installation, run databricks --version before anything else to confirm the binary is in your PATH and meets the version floor

More Patterns

Linux Installation

“Install the Databricks CLI on a Linux machine or CI runner using the official install script.”

# System-wide install (requires sudo)
curl -fsSL https://raw.githubusercontent.com/databricks/setup-cli/main/install.sh | sh

# Verify
databricks --version

The install script detects your architecture (amd64/arm64) and downloads the correct binary. On headless CI runners, this is the fastest path to a working CLI.

User-Directory Install (No Sudo)

“Install the CLI in my home directory on a sandboxed environment where I don’t have admin access.”

# Install to ~/.local/bin (no sudo required)
curl -fsSL https://raw.githubusercontent.com/databricks/setup-cli/main/install.sh \
  | sh -s -- --install-dir ~/.local/bin

# Make sure ~/.local/bin is in your PATH
export PATH="$HOME/.local/bin:$PATH"

# Verify
databricks --version

This is the right approach for IDE containers, Codespaces, or shared servers where you don’t have root access. Add the export PATH line to your ~/.bashrc or ~/.zshrc so it persists.

Windows Installation

“Install the Databricks CLI on Windows using WinGet.”

winget install Databricks.DatabricksCLI

# Verify
databricks --version

WinGet handles PATH registration automatically. After installation, open a new terminal window for the databricks command to be recognized.

Upgrading to the Latest Version

“Upgrade the CLI to the latest version to pick up new bundle features.”

# macOS
brew upgrade databricks

# Linux (re-run the install script)
curl -fsSL https://raw.githubusercontent.com/databricks/setup-cli/main/install.sh | sh

# Verify the upgrade
databricks --version

New CLI releases ship roughly every two weeks. If a bundle feature isn’t available or a command returns an unexpected error, upgrading the CLI is the first troubleshooting step.

Watch Out For

• Running an old CLI version — Bundle features require v0.292.0 or later. If databricks bundle validate returns “unknown command,” you’re on an outdated version. Upgrade first.
• Forgetting to re-run the install script on Linux — Unlike Homebrew, the Linux install script doesn’t auto-update. You get whatever version was current when you last ran it. Re-run the script periodically or pin it in your CI setup.
• PATH conflicts with older installations — If you previously installed via pip install databricks-cli (the old Python-based CLI), the old binary may shadow the new one. Run which databricks to confirm you’re using the Go-based CLI at the expected path.
• Missing PATH entry in containers — The --install-dir flag puts the binary where you ask, but doesn’t modify your shell profile. If databricks --version returns “command not found” after installation, add the install directory to $PATH in your shell config.