Getting Started¶
This guide will help you get started with the YAML Workflow Engine.
Installation¶
Option 1: pipx (recommended for CLI use)
pipx installs the yaml-workflow command in an isolated environment, so it never conflicts with your project's dependencies:
# Core CLI (run, validate, visualize, init)
pipx install yaml-workflow
# With web dashboard
pipx install 'yaml-workflow[serve]'
# With MCP server (AI agent integration)
pipx install 'yaml-workflow[mcp]'
# Everything
pipx install 'yaml-workflow[all]'
Install pipx first if needed: brew install pipx (macOS) or pip install pipx.
Option 2: pip
Install into the current Python environment (useful when using yaml-workflow as a library):
pip install yaml-workflow # Core
pip install 'yaml-workflow[serve]' # + web dashboard
pip install 'yaml-workflow[mcp]' # + MCP server
pip install 'yaml-workflow[all]' # Everything
Option 3: Docker
Run without installing Python:
Basic Concepts¶
The YAML Workflow Engine is built around a few core concepts:
- Workflows: YAML files that define a sequence of tasks to be executed
- Tasks: Individual units of work that can be executed
- Flows: Named sequences of tasks that can be executed together
- Parameters: Values that can be passed to workflows and tasks
Your First Workflow¶
- Create a new directory for your workflow:
- Initialize a new workflow project:
- Examine the generated workflow file (
workflows/hello_world.yaml):
name: Hello World Workflow
description: A simple example workflow
params:
name:
description: Name to include in greeting
type: string
required: true
steps:
- name: greet
task: shell
command: echo "Hello, {{ name }}!"
- Run the workflow:
Next Steps¶
- Learn about workflow configuration
- Explore built-in tasks
- See more examples