Skip to main content
The ZenML CLI is the primary interface for interacting with ZenML from your terminal. It provides commands to manage pipelines, stacks, integrations, and more.

​
Getting Started

Check your ZenML version: 
zenml version
Get help on any command: 
zenml --help
zenml <command> --help
zenml <command> <subcommand> --help

​
Command Categories

ZenML CLI commands are organized into several categories:

Pipeline Commands

Manage pipelines, runs, builds, and deployments

Stack Commands

Configure and manage your MLOps stacks

Integration Commands

Install and manage third-party integrations

Deployment Commands

Deploy and manage pipeline deployments

​
Common Patterns

​
Using the Help Flag

Every command supports the --help flag to display detailed usage information: 
zenml artifact-store register --help

​
Understanding Concepts

Use the explain subcommand to learn about ZenML concepts: 
zenml artifact-store explain
zenml orchestrator explain

​
Filtering List Commands

Many list commands support powerful filtering options. Use --help to see available filters: 
zenml orchestrator list --help

​
Filter by Text Fields

For TEXT or UUID fields, use these operators: 
# Contains
zenml orchestrator list --name "contains:sagemaker"

# Starts with
zenml pipeline list --name "startswith:training"

# Ends with
zenml stack list --name "endswith:prod"

​
Filter by Boolean Fields

For BOOL fields, use True or False: 
zenml stack list --is_shared True

​
Filter by Date Fields

For DATETIME fields, use comparison operators with the format YYYY-MM-DD HH:MM:SS: 
# Greater than
zenml orchestrator list --created "gt:2021-01-01 00:00:00"

# Less than or equal
zenml pipeline runs list --created "lte:2024-12-31 23:59:59"

# Greater than or equal
zenml stack list --updated "gte:2024-01-01 00:00:00"

​
Combine Filters

Filters can be combined using and and or keywords: 
zenml pipeline runs list --status "running or completed" --created "gte:2024-01-01 00:00:00"

​
Sorting Results

Use the --sort_by option to sort list command output: 
# Ascending order
zenml orchestrator list --sort_by "asc:name"

# Descending order
zenml pipeline runs list --sort_by "desc:created"

​
Output Formats

All list commands support multiple output formats via the --output or -o flag: 
# JSON format (for programmatic use)
zenml stack list --output=json | jq '.items[] | select(.name=="production")'

# CSV format (for spreadsheet analysis)
zenml pipeline runs list --output=csv > pipeline_runs.csv

# YAML format (for configuration)
zenml deployment list --output=yaml

# TSV format (for simpler parsing)
zenml stack list --output=tsv

# Table format (default, human-readable)
zenml pipeline list --output=table

​
Custom Column Selection

Control which columns to display: 
zenml stack list --columns=id,name,orchestrator
zenml pipeline list --columns=id,name --output=json

​
Environment Variables

Set default output preferences: 
# Set default output format
export ZENML_DEFAULT_OUTPUT=json

# Set consistent table width
export ZENML_CLI_COLUMN_WIDTH=120

​
Global Options

These options work with most ZenML commands:
OptionDescription
--helpShow help message and exit
--version, -vShow ZenML version
--debugEnable debug logging
--no-colorDisable colored output

​
Working Without a Repository

Some commands (like zenml pipeline register or zenml pipeline run) can be executed outside a ZenML repository. When you do this:
  • ZenML uses your current working directory as the source root
  • You’ll see a warning message suggesting to run zenml init
  • The command will still work, but relative imports are resolved from your current directory
For best results, run zenml init at your project’s source code root before using ZenML commands.

​
Component Labels

Many stack components support labels for organization and filtering: 
# Add labels when registering
zenml artifact-store register my-store --flavor=s3 -l environment=prod -l team=ml

# Filter by labels
zenml artifact-store list --label environment=prod

​
Next Steps

Initialize Project

Set up a new ZenML repository

Login to Server

Connect to ZenML Pro or a self-hosted server

Configure Stacks

Learn about stack management

Run Pipelines

Execute and manage pipeline runs

Build docs developers (and LLMs) love

Get started for freeTalk to us