CLI Configuration

The buddy CLI stores its configuration and credentials on your machine and manages them through a handful of commands.

Where config lives

Configuration is kept under ~/.config/buddy/:

File	Purpose
`config.json`	User info, active workspace, and workspace map
`credentials.json`	Saved auth token

These paths come from BuddyCliSettings (buddy.cli.settings) and can be overridden with BUDDY_-prefixed environment variables.

Setting up

buddy setup        # set up Buddy and log in to the platform
buddy init         # initialize the Buddy config
buddy init -l      # initialize and log in

buddy setup runs initialize_buddy(login=True), which prepares the config and starts the browser login flow. buddy init initializes the config; add -l/--login to authenticate at the same time.

init configures Buddy, not a project

buddy init sets up Buddy's own configuration — it does not create a project folder. Use buddy ws create to scaffold a workspace.

Authentication

Logging in opens your browser to the Buddy sign-in page; once you approve, a token is posted back to a short-lived local server and saved to credentials.json. If the browser flow isn't available, set an API key via environment variable instead:

export BUDDY_API_KEY=your_api_key   # PowerShell: $env:BUDDY_API_KEY="..."

Check connectivity and that you're authenticated:

buddy ping

Viewing config

buddy config       # print user, active workspace, and all workspaces

Resetting

buddy reset        # reset the existing configuration
buddy init -r      # equivalent: reset during init

Resetting removes the local Buddy configuration directory so you can start fresh.

Environment variables

buddy reads BUDDY_-prefixed variables (via pydantic-settings). Common ones:

Variable	Effect
`BUDDY_API_KEY`	Authenticate without the browser flow
`BUDDY_API_RUNTIME`	Target runtime: `dev`, `stg`, or `prd`

Provider keys (OPENAI_API_KEY, ANTHROPIC_API_KEY, …) are read by the model clients, not the CLI config. See .env.example for the full list.