Using Looker CLI (looker-cli)

This skill documents standard operations, session management, discovery workflows, and API payload tools in Looker CLI (looker-cli).

1. Managing Profiles

Profiles allow you to store connection configurations (host, port, credentials, tokens) for different Looker instances (e.g., development, staging, production) in ~/.config/looker-cli/config.yaml.

Adding a Profile

Add connection settings for a Looker instance using profile add:

looker-cli profile add <profile_name> --host <instance_host> --port <port>

Optional flags when creating a profile:

• --client-id <id>: Set Looker API Client ID.
• --client-secret <secret>: Set Looker API Client Secret.
• --token <access_token>: Store a pre-acquired API access token.
• --refresh-token <refresh_token>: Store an OAuth refresh token.

Example:

looker-cli profile add staging --host staging.looker.mycompany.com --port 443

Listing Profiles

List all configured profiles:

looker-cli profile ls

The active (default) profile is denoted with an asterisk (*):

  dev (dev.looker.mycompany.com:443)
* staging (staging.looker.mycompany.com:443)
  prod (prod.looker.mycompany.com:443)

Setting the Default Profile

Set a specific profile as your default active profile using profile use:

looker-cli profile use <profile_name>

Example:

looker-cli profile use prod

Subsequent commands will automatically execute against the active profile without needing explicit connection flags.

Overriding Profile per Command (--profile)

To execute a single command against a profile other than the default active profile, use the global --profile flag:

looker-cli <command> --profile <profile_name>

Examples:

# Query current user on production without changing active default profile
looker-cli user me --profile prod

# List projects in dev environment
looker-cli project ls --profile dev

Deleting a Profile

Remove a stored profile definition using profile rm:

looker-cli profile rm <profile_name>

2. Persistent Session Login (session login)

looker-cli supports persistent authentication sessions, allowing commands to run without authenticating on every request.

Basic Persistent Login

Authenticate against the host defined in your active profile using configured API credentials:

looker-cli session login

Upon successful authentication, access credentials and session state are persisted to ~/.config/looker-cli/config.yaml.

OAuth 2.0 PKCE Browser Login (--oauth)

For user-scoped interactive logins using OAuth 2.0 PKCE, pass the --oauth flag:

looker-cli session login --oauth

Workflow:

1. The CLI launches an authorization URL (or outputs the URL in headless environments).
2. Complete authentication and grant access in the browser.
3. OAuth access tokens and refresh tokens are automatically saved to your active profile in config.yaml.

Outputting Token to Terminal (--text)

To display the authenticated access token to stdout (e.g., for piping into scripts or environment variables) instead of persisting it to the config file, pass --text:

looker-cli session login --text

Checking Active Session Status

View details about the currently active session (authenticated user ID, session workspace, expiration):

looker-cli session get

Workspace Context Switching

Switch your persistent session context between personal development workspace (dev) and production (production):

# Switch to personal development mode
looker-cli session update --workspace-id dev

# Switch to production mode
looker-cli session update --workspace-id production

Ending a Session

Invalidate and terminate the active persistent session:

looker-cli session logout

3. Command Discovery with Meta Utilities

looker-cli includes meta discovery utilities to inspect command hierarchies and search available API endpoints.

Visualizing Command Subtrees (meta tree)

Display the interactive CLI command tree:

looker-cli meta tree

Scoping to a Specific Noun (--noun)

To filter the tree for subcommands pertaining to a specific resource noun (e.g., project, dashboard, user, query), pass --noun:

looker-cli meta tree --noun project

Sample output:

project
├── branch
├── cat
├── checkout
├── create
├── deploy
│   └── key
├── directory
│   ├── create
│   ├── ls
│   └── rm
├── file
│   ├── cat
│   ├── create
│   ├── ls
│   ├── rm
│   └── update
├── import
├── ls
├── update
└── validate

JSON Output Format (--output json)

Output the command structure as a JSON object:

looker-cli meta tree --noun query --output json

Searching Commands by Keyword (meta search)

Search all high-level CLI commands and low-level API operations by keyword:

looker-cli meta search <KEYWORD>

Examples:

# Search for commands related to projects
looker-cli meta search project

# Search for commands related to users or authentication
looker-cli meta search user

Sample output:

Found 63 matching commands:
  looker-cli api project create_project - Create Project
  looker-cli api project deploy_to_production - Deploy To Production
  looker-cli api project validate_project - Validate Project
  looker-cli project create - Create a new project
  looker-cli project validate - Validate a project

4. Inspecting API Request Bodies (--describe-body & --template)

looker-cli exposes low-level REST API endpoints under looker-cli api <group> <endpoint>. When creating or updating Looker resources via raw API commands, use --describe-body and --template to inspect expected parameters and structure payload files.

Inspecting Request JSON Schema (--describe-body)

Pass --describe-body on an API command to output the formal JSON Schema of its request body. This lists all allowable parameters, data types, field descriptions, write-only flags, and allowed values (x-looker-values):

looker-cli api <group> <endpoint> --describe-body

Example: Inspect schema for creating a project

looker-cli api project create_project --describe-body

Sample snippet:

{
  "properties": {
    "name": {
      "description": "Project display name",
      "type": "string",
      "x-looker-nullable": false
    },
    "pull_request_mode": {
      "description": "The git pull request policy for this project. Valid values are: \"off\", \"links\", \"recommended\", \"required\".",
      "type": "string",
      "x-looker-nullable": false,
      "x-looker-values": [
        "off",
        "links",
        "recommended",
        "required"
      ]
    }
  }
}

Generating Payload Skeleton Templates (--template)

Pass --template on an API command to print a clean JSON skeleton populated with default/empty values for writable fields:

looker-cli api <group> <endpoint> --template

Example: Generate template for creating a connection

looker-cli api connection create_connection --template

Sample output:

{
  "database": "",
  "db_timezone": "",
  "dialect_name": "",
  "host": "",
  "name": "",
  "port": ""
}

Practical Workflow: Creating Resources via API Payloads

Combine --template with piping or file redirection to construct and submit POST/PUT payloads:

1. Generate and edit template file:

``bash looker-cli api project createproject --template > newproject.json ``

1. Populate required fields in new_project.json:

``json { "name": "mynewproject", "pullrequestmode": "recommended" } ``

1. Execute API request using payload file:

``bash looker-cli api project createproject --body newproject.json ``

5. Quick Reference Cheat Sheet