How to install and configure the dbt Cloud CLI

The dbt Cloud CLI, now called the dbt platform CLI, lets you write code in your own editor and run dbt commands from your terminal, while the commands execute on dbt’s infrastructure with the connection and credentials of your development environment. There is no profiles.yml to maintain: the CLI signs in with a personal access token kept in dbt_cloud.yml.

Copy and paste

1. Install it with Homebrew:

which dbt                      # should print "dbt not found"; see below if it does not
brew untap dbt-labs/dbt
brew tap dbt-labs/dbt-cli
brew install dbt
dbt --help

The first line of the help should read The dbt CLI - an ELT tool for running SQL transformations and data models in dbt....

2. Download your credentials. In dbt platform, open Account settings > Your profile > CLI > Configure Cloud authentication and click Download CLI configuration file. Save it as:

~/.dbt/dbt_cloud.yml

3. Link your local project to the dbt platform project, in dbt_project.yml:

dbt-cloud:
  project-id: '123456'

4. Check the connection from the project folder:

dbt environment show
dbt compile

Before you start

• The project must have a development environment in dbt platform.
• Your development credentials for the project must be set in Account settings. The CLI uses them for the warehouse connection.
• Clone the project’s repository on your machine. The CLI works inside that folder.

dbt_cloud.yml

The downloaded file holds the account, the project and your personal access token:

version: "1"
context:
  active-host: "ab123.us1.dbt.com"
  active-project: "123456"
projects:
  - project-name: "Shop"
    project-id: "123456"
    account-name: "Example"
    account-id: "98765"
    account-host: "ab123.us1.dbt.com"
    token-name: "cloud-cli-1234"
    token-value: "dbtu_..."

The CLI looks for exactly ~/.dbt/dbt_cloud.yml. A browser that saves it as dbt_cloud(2).yml breaks the setup until you rename it. The token gives access to your account: do not commit the file.

Next to dbt Core

The CLI installs a command called dbt, like dbt Core and dbt v2, and they conflict when both are on the PATH.

• dbt Core installed globally: uninstall it with pip uninstall dbt-core, or use the CLI only.
• dbt Core in a virtual environment: deactivate the environment before using the CLI, or give the CLI an alias:

alias dbt-cloud="/opt/homebrew/bin/dbt"

• pip install dbt is not the CLI. On PyPI, the dbt package now installs dbt v2. The platform CLI comes from Homebrew.

What is different from dbt Core

• Deferral is automatic. Models you do not build are read from production; --no-defer turns it off. See how to defer to production.
• Artifacts are downloaded to your target folder after each command; --download-artifacts=false skips that.
• Sessions. dbt invocation list shows running commands, and dbt reattach reconnects to one.
• Linting. dbt sqlfluff lint, fix and format are built in.
• Local packages. Relative paths in packages.yml are not supported.

Updating

brew update && brew upgrade dbt

Based on the dbt platform documentation (October 2026); not run against a live dbt account.