Skip to main content

Overview

ClawKitchen’s workflow system provides a complete visual interface for creating, managing, and operating automation workflows. Built on the file-first ClawRecipes foundation, it combines ease-of-use with the power and flexibility of code-based workflows.

Core Concepts

File-First Architecture

Workflows in ClawKitchen are real operating artifacts stored as files in your team workspace:
  • Portable: Workflows can be version controlled, shared, and backed up
  • Inspectable: Direct file access for debugging and understanding
  • Durable: Workflows persist independently of the UI
  • Collaborative: Team members can edit workflows through git

Visual Editor

The workflow visual editor provides:
  • Drag-and-Drop Design: Build workflows without writing JSON
  • Node Palette: Pre-built components for common workflow tasks
  • Real-Time Validation: Immediate feedback on workflow configuration
  • Template System: Start with proven workflow patterns

Execution Engine

Workflows are executed by the ClawRecipes runtime:
  • Distributed: Nodes can run on different agents
  • Resilient: Automatic retry and error handling
  • Observable: Complete audit trails and debugging information
  • Scalable: Handle complex multi-step automations

Workflow Components

Node Types

Control Nodes

  • Start: Entry point for workflow execution
  • End: Explicit workflow completion point

Logic Nodes

  • LLM: AI-powered content generation and decision making
  • Human Approval: Manual review and approval steps

Action Nodes

  • Tool: Execute commands, send messages, manipulate files
  • Writeback: Record workflow results in team files

Media Nodes

  • Media-Image: Generate images from text prompts
  • Media-Video: Create video content automatically
  • Media-Audio: Generate audio and voiceover content

Edge Types

  • Success: Execute when previous node succeeds
  • Error: Handle failure cases and errors
  • Always: Execute regardless of previous node outcome

Advanced Features

Template Variables

Dynamic content insertion throughout workflows:
  • Run Context: {{run.id}}, {{workflow.id}}, {{date}}
  • Node Outputs: {{previous_node.result}}, {{draft_content.title}}
  • Team Data: {{team.name}}, {{agent.role}}

Memory Integration

Automatic team memory injection into LLM nodes:
  • Context Awareness: Workflows understand team knowledge and procedures
  • Consistent Outputs: AI responses align with team standards and decisions
  • Learning: Workflows improve based on accumulated team experience

Approval Workflows

Human oversight and control:
  • Configurable Bindings: Connect to Slack, Discord, email, or custom channels
  • Rich Context: Approvers see full workflow context and generated content
  • Revision Cycles: Support for feedback and iteration

Getting Started

Creating Your First Workflow

  1. Navigate to Workflows
    • Open your team workspace
    • Click the “Workflows” tab
    • View existing workflows or create new ones
  2. Choose Creation Method
    • From Template: Start with proven patterns (recommended)
    • From Scratch: Build completely custom workflows
    • Import: Load workflows from files or other teams
  3. Use the Visual Editor
    • Drag nodes from the palette onto the canvas
    • Connect nodes to define execution flow
    • Configure each node’s properties and settings
    • Test with “Run Now” functionality

Marketing Content Pipeline

Complete content creation and distribution:
Draft Content → Generate Visuals → Human Review → Publish to Platforms

Documentation Maintenance

Keep documentation current with code changes:
Detect Changes → Draft Updates → Review → Update Files → Notify Team

Customer Support Automation

Intelligent ticket routing and response:
Analyze Ticket → Route by Priority → Generate Response → Human Handoff

Workflow Operations

Triggering Workflows

Manual Execution

  • Run Now: Immediate workflow execution for testing
  • Scheduled Runs: One-time execution at specified time
  • Bulk Operations: Run workflows across multiple contexts

Automated Triggers

  • Cron Schedules: Time-based workflow execution
  • Event Triggers: React to external events or changes
  • Chain Execution: Workflows that trigger other workflows

Monitoring and Debugging

Run Inspection

Comprehensive visibility into workflow execution:
  • Live Status: Real-time workflow progress
  • Node Outputs: Generated content and data at each step
  • Error Details: Complete error context and stack traces
  • Performance Metrics: Execution time and resource usage

Approval Management

Streamlined approval workflows:
  • Pending Approvals: Dashboard of items awaiting review
  • Approval History: Complete audit trail of decisions
  • Bulk Actions: Approve or reject multiple items
  • Delegation: Route approvals to appropriate team members

Workflow Collaboration

Team Editing

  • Concurrent Access: Multiple team members can edit workflows safely
  • Change Tracking: Git integration for version control
  • Permission Management: Control who can edit vs. run workflows
  • Review Process: Optional approval for workflow changes

Knowledge Sharing

  • Workflow Libraries: Share successful patterns across teams
  • Documentation: Rich descriptions and usage guides
  • Template Creation: Convert working workflows into reusable templates
  • Best Practices: Team conventions and standards

Advanced Capabilities

Integration Ecosystem

Tool Integration

Extensive tool ecosystem support:
  • Native Tools: Built-in support for common operations
  • Skill System: Extensible through OpenClaw skills
  • API Connectors: Connect to external services and APIs
  • Custom Tools: Build team-specific automation tools

Media Generation

AI-powered content creation:
  • Multi-Provider: Support for multiple generation services
  • Auto-Discovery: Automatic detection of available capabilities
  • Quality Control: Human review and iteration cycles
  • Asset Management: Organized storage and reuse of generated content

External Publishing

  • Social Media: Direct publishing to Twitter, LinkedIn, Facebook
  • Documentation: Update wikis, knowledge bases, and documentation
  • Communication: Send notifications and updates to team channels
  • Data Export: Push results to databases and external systems

Performance and Scalability

Optimization

  • Parallel Execution: Run independent nodes simultaneously
  • Caching: Reuse expensive computations and API calls
  • Resource Management: Control memory and compute usage
  • Rate Limiting: Respect external service limits

Monitoring

  • Metrics Dashboard: Track workflow performance and success rates
  • Alerting: Notifications for failures and performance issues
  • Logging: Comprehensive logs for debugging and auditing
  • Analytics: Usage patterns and optimization opportunities

Best Practices

Workflow Design

  • Start Simple: Begin with linear workflows before adding complexity
  • Error Handling: Include error paths for all critical operations
  • Modularity: Break complex workflows into reusable components
  • Documentation: Clearly describe workflow purpose and usage

Team Collaboration

  • Naming Conventions: Consistent workflow and node naming
  • Testing: Thorough testing before deploying to production
  • Version Control: Use git branches for experimental changes
  • Knowledge Transfer: Document team-specific patterns and decisions

Security and Governance

  • Access Control: Appropriate permissions for editing and execution
  • Approval Gates: Human oversight for sensitive operations
  • Audit Trails: Complete records of workflow changes and executions
  • Secret Management: Secure handling of API keys and credentials

Troubleshooting

Common Issues

Workflow Won’t Start
  • Verify all required agents are available
  • Check workflow trigger configuration
  • Ensure necessary plugins are installed
  • Review team permissions and access
Node Execution Failures
  • Check node configuration and parameters
  • Verify tool availability and permissions
  • Review error logs in run detail pages
  • Test node configuration independently
Approval Timeouts
  • Verify approval binding configuration
  • Check notification delivery to reviewers
  • Review approval channel connectivity
  • Consider timeout value adjustments

Getting Help

  • Run Logs: Detailed execution information
  • Team Memory: Documented procedures and troubleshooting guides
  • Template Examples: Reference implementations for common patterns
  • Community Resources: Shared workflows and best practices
For deeper technical details, see: Workflows in ClawKitchen transform manual processes into reliable, observable automation that scales with your team’s needs while maintaining human oversight and control.