Quick Start

Get up and running with WireFlow in 5 minutes. This guide assumes you've already installed WireFlow and configured your API key.

Your First Project (5 Minutes)

Let's create a simple project that uses AI to analyze a text file.

Step 1: Initialize a Project

Create a new directory and initialize it as a Workflow project:

mkdir my-analysis
cd my-analysis
wfw init .

This creates a .workflow/ directory with default configuration.

Step 2: Create Sample Content

Let's create a file to analyze:

cat > notes.txt << 'EOF'
Meeting Notes - Project Planning

Key Discussion Points:
- Need to refactor the authentication module
- API rate limiting causing issues in production
- Documentation is outdated
- Team wants to adopt TypeScript

Action Items:
- Review authentication code for security issues
- Implement exponential backoff for API calls
- Update README with new configuration options
- Create migration plan for TypeScript

Timeline: Complete by end of Q2
EOF

Step 3: Create Your First Workflow

Create a workflow to extract action items:

wfw new extract-actions

This creates .workflow/run/extract-actions/ with:

• task.txt - Your prompt/task description
• config - Workflow-specific configuration

Step 4: Define the Task

Edit the task file (opens in your $EDITOR):

wfw edit extract-actions

Or edit it manually:

cat > .workflow/run/extract-actions/task.txt << 'EOF'
Extract all action items from the meeting notes and format them as a checklist.
For each item, identify:
1. What needs to be done
2. Any relevant deadlines or priorities
3. Dependencies on other tasks

Format the output as a markdown checklist with clear priorities.
EOF

Step 5: Add Context

Add your notes file as context:

wfw run extract-actions -cx notes.txt --stream

Streaming Mode

The --stream flag enables real-time output as the API generates the response. Omit it for buffered mode (response saved to file after completion).

Step 6: View the Output

The response is saved to .workflow/run/extract-actions/output.md. Use the cat subcommand for easy viewing:

wfw cat extract-actions

Or access the file directly:

cat .workflow/run/extract-actions/output.md

What Just Happened?

Let's break down what you did:

1. wfw init . - Created a .workflow/ project structure
2. wfw new extract-actions - Created a new workflow subdirectory
3. wfw edit extract-actions - Opened the task prompt for editing
4. wfw run extract-actions - Executed the workflow with the Anthropic API

The workflow:

• Read your task from task.txt
• Gathered context from notes.txt
• Sent everything to Claude
• Saved the response to .workflow/run/extract-actions/output.md
• Kept a backup of any previous output

Key Concepts

Project Structure

my-analysis/
├── notes.txt                  # Your content
└── .workflow/                 # Wireflow project root
    ├── config                 # Project-level config
    ├── output/                # Hardlinks to outputs
    └── run/                   # Workflow directories
        └── extract-actions/   # Individual workflow
            ├── task.txt       # Task/prompt
            ├── config         # Workflow config
            └── output.md      # Generated output

Modes of Operation

Workflow Mode (what you just used):

• Creates persistent directory structure
• Stores task, config, and outputs
• Ideal for iterative development

Task Mode (quick one-offs):

wfw task -i "Summarize these notes" -cx notes.txt

• No persistent directories
• Output streams to stdout
• Great for quick queries

Try These Next

Run with More Context

Use glob patterns to add multiple files:

wfw run extract-actions -cx "*.txt" --stream

Modify the Configuration

Change the AI model or parameters:

# View current config
wfw config extract-actions

# Edit workflow config
wfw edit extract-actions

Add configuration like:

MODEL=claude-opus-4-5-20251101
TEMPERATURE=0.7
MAX_TOKENS=2000

Create a Dependent Workflow

Create a second workflow that uses the first's output:

# Create new workflow
wfw new prioritize-actions

# Edit task
wfw edit prioritize-actions
# Add task: "Take these action items and create a prioritized implementation plan"

# Run with dependency
wfw run prioritize-actions --depends-on extract-actions --stream

The --depends-on flag automatically includes the output of extract-actions as context!

Try Task Mode

For quick queries without creating workflows:

# Inline task
wfw task -i "What are the main themes in these notes?" -cx notes.txt

# Named task (create reusable task templates)
mkdir -p ~/.config/wireflow/tasks
echo "Summarize the key points in bullet format" > ~/.config/wireflow/tasks/summarize.txt
wfw task summarize -cx notes.txt

Common Commands

# List all workflows in project
wfw list

# View configuration cascade
wfw config extract-actions

# Get help
wfw help
wfw help run
wfw run -h

Next Steps

You now know the basics! Here's what to explore next:

1. First Workflow Tutorial - Detailed walkthrough of a real-world workflow
2. Configuration Guide - Master the configuration cascade
3. Execution Modes - Learn all the options for running workflows and tasks

Quick Reference

Command	What It Does
wfw init <dir>	Initialize a workflow project
wfw new <name>	Create a new workflow
wfw edit <name>	Edit workflow task
wfw config <name>	View configuration
wfw run <name>	Execute a workflow
wfw task -i "<text>"	Quick one-off task
wfw cat <name>	Display workflow output
wfw open <name>	Open workflow output in default app
wfw list	List all workflows
wfw help	Show help

---

Ready for a more detailed tutorial? Continue to First Workflow →