# The Databricks CLI wraps the REST API for your terminal

> What the Databricks CLI is, how it maps onto the REST API, and which version numbers are generally available, preview or legacy.

Feature: Databricks CLI. Release stage: Generally available, as of 21 May 2026 (https://docs.databricks.com/aws/en/release-notes/product/2026/may#databricks-cli-is-now-ga). Documentation: https://docs.databricks.com/aws/en/dev-tools/cli
By Raja Vemuri, 8 October 2026. Databricks Field Guide, TechFabric. Canonical: https://databricks.techfabric.com/features/databricks-cli/

The Databricks CLI is a command-line interface that lets you work with the Databricks platform from your local terminal or from automation scripts. It wraps the Databricks REST API, which provides endpoints for modifying or requesting information about account and workspace objects, so anything you can do with an API call you can do with a command. Databricks' release notes record that the Databricks CLI reached general availability on 21 May 2026. You can also run its commands inside a workspace through the web terminal.

It ships as a dependency-free binary, and the source is public in the `databricks/cli` GitHub repository.

## How it works

Each command maps onto a REST endpoint. Here is how you print information about an individual cluster in a workspace.

```bash
databricks clusters get 1234-567890-a12bcde3
```

The documentation gives the equivalent `curl` call, which shows what the CLI is doing underneath.

```bash
curl --request GET "https://${DATABRICKS_HOST}/api/2.0/clusters/get" \
     --header "Authorization: Bearer ${DATABRICKS_TOKEN}" \
     --data '{ "cluster_id": "1234-567890-a12bcde3" }'
```

The host and token come from configuration instead of being typed each time. Databricks documents several authentication types between the CLI and your accounts and workspaces, and configuration profiles let you switch between related groups of settings without re-authenticating. To submit and manage GPU training workloads, the `databricks air` commands cover the AI Runtime.

## Version release types

The release type of the Databricks CLI corresponds to its version number, which is the part people get wrong when they inherit a machine with an old install:

```bash
databricks --version
```

Versions 1.0.0 and above are generally available. Versions 0.205 through 0.299 are Public Preview. Versions 0.18 and below are the legacy CLI, for which Databricks plans no support and no new feature work. Databricks publishes a migration guide for moving from version 0.18 or below to 0.205 or above, and that migration is not a drop-in swap, because the command surface changed.

## Limits and availability

Databricks CLI use is subject to the Databricks License and the Databricks Privacy Notice, including any Usage Data provisions. The documentation page was last updated on 24 September 2026. The general availability milestone is recorded in the May 2026 product release notes, dated 21 May 2026, under "Databricks CLI is now GA". Installation is covered by the install guide, and a Homebrew tap and a container image at `ghcr.io/databricks/cli` are available from the project repository.

## When to use it

Use it anywhere a human would otherwise click. Creating workspaces, deploying Asset Bundles and inspecting a job run from a build agent all work the same way from a script as from a browser. The version rule matters more than it looks, because a 0.18 install still runs and still accepts some old commands, so a CI job can quietly be using an unsupported tool. Pin the CLI version in your build image the way you pin any other dependency, and check it the first time a pipeline does something strange. [CI/CD and Environments](/guide/ci-cd/) covers the deployment side of this, and [Accounts and Workspaces](/guide/accounts-and-workspaces/) covers the infrastructure-as-code habits the CLI supports.

## Questions people ask

### What is the Databricks CLI?

It is a command-line interface for interacting with the Databricks platform from a local terminal or an automation script. It wraps the Databricks REST API, so each command corresponds to an endpoint that modifies or reads account and workspace objects.

### Is the Databricks CLI generally available?

Yes. Databricks' May 2026 release notes record that the Databricks CLI went GA on 21 May 2026, and versions 1.0.0 and above are the generally available releases.

### Which Databricks CLI version is the legacy one?

Versions 0.18 and below are the legacy Databricks CLI, and Databricks plans no support or new feature work for it. Versions 0.205 through 0.299 are Public Preview, and Databricks publishes a migration path from 0.18 or below to 0.205 or above.