Command-Line Interface

Overview

The Arize Toolkit CLI lets you manage models, monitors, prompts, evaluators, dashboards, traces, and more directly from your terminal — no Python scripting required.

Every command maps 1-to-1 to a method on the Python Client, so anything you can do in code you can also do from the command line.

Key features:

1. Rich table output by default, with a --json flag for machine-readable JSON
2. Configuration profiles stored in ~/.arize_toolkit/config.toml (similar to AWS CLI profiles)
3. projects alias — arize_toolkit projects list and arize_toolkit models list are interchangeable

Installation

The CLI is an optional extra. Install it alongside the core library:

pip install arize_toolkit[cli]

This pulls in Click for command parsing, Rich for terminal output, and TOML libraries for profile management.

Quick Setup

1. Create a configuration profile

arize_toolkit config init

You will be prompted for your Developer API Key, organization name, space name, and app URL. The result is saved to ~/.arize_toolkit/config.toml.

2. Verify the connection

arize_toolkit spaces list

If everything is configured correctly you will see a table of spaces in your organization.

3. Try a few commands

# List all models / projects
arize_toolkit models list

# Get details for a specific model (JSON output)
arize_toolkit --json models get "fraud-detection-v3"

# List monitors filtered by category
arize_toolkit monitors list --model-name "fraud-detection-v3" --category performance

# List all prompts
arize_toolkit prompts list

Global Options

Every command accepts the following flags. They can appear before the subcommand:

Flag	Description
--profile NAME	Use a named configuration profile instead of default
--json	Output raw JSON instead of Rich tables
--api-key KEY	Override the API key for this invocation
--org NAME	Override the organization name
--space NAME	Override the space name
--app-url URL	Override the Arize app URL
--help	Show help for any command or group

Configuration resolution order (highest priority first):

1. CLI flags (--api-key, --org, --space)
2. Environment variables (ARIZE_DEVELOPER_KEY, ARIZE_ORGANIZATION, ARIZE_SPACE)
3. Profile from ~/.arize_toolkit/config.toml

Command Groups

Group	Description	Page
config	Manage configuration profiles	Configuration
spaces	List, create, and switch spaces	Spaces
orgs	List and create organizations	Organizations
users	Search users and manage space membership	Users
models / projects	List models, check volume, pull performance metrics	Models & Projects
monitors	Create, list, copy, and delete monitors	Monitors
prompts	Manage prompt templates and versions	Prompts
custom-metrics	Create and manage custom metrics	Custom Metrics
evaluators	Manage LLM and code evaluators	Evaluators
dashboards	Create and view dashboards	Dashboards
traces	List traces, get spans, and discover available columns	Traces
imports	Manage file and table import jobs	Data Imports

JSON Mode

Pass --json before the subcommand to get raw JSON output. This is useful for piping results to jq or other tools:

# Pretty-print all model names
arize_toolkit --json models list | jq '.[].name'

# Store monitor details in a file
arize_toolkit --json monitors get "accuracy-monitor" --model "fraud-detection-v3" > monitor.json

Profiles

Profiles let you maintain separate configurations for different environments:

# Create a profile for staging
arize_toolkit config init --profile-name staging

# Use it for a single command
arize_toolkit --profile staging models list

# Switch the default profile
arize_toolkit config use staging

See Configuration for full details.