Creating Workflows

Learn how to build workflows from scratch using the visual workflow builder. This guide walks you through each step of the workflow creation process.

Step 1: Navigate to Workflow Builder

1. Click Workflows in the main navigation
2. Click Create Workflow button
3. You'll be taken to the visual workflow builder canvas

The workflow builder provides:

• A canvas for arranging nodes
• A node palette on the left with all available components
• A configuration panel on the right for node settings
• A toolbar with test, deploy, and save options

Step 2: Add a Trigger Node

Every workflow starts with a trigger. The trigger determines when and how your workflow executes.

Selecting a Trigger

From the node palette on the left, select one of these trigger types:

Trigger Options

• Webhook: Trigger from external HTTP requests with HMAC signature verification
• Schedule: Run on interval (minutes), daily at specific time, or custom cron expression
• REST API: Expose workflow as authenticated API endpoint
• Form: Trigger from form submissions with CAPTCHA and validation

Configuring Your Trigger

Click the trigger node to open its configuration panel. Each trigger type has specific settings:

Webhook Trigger

• Generate webhook URL
• Configure HMAC signature verification
• Set allowed HTTP methods
• Define request validation rules

Schedule Trigger

• Set interval (every X minutes)
• Daily execution time
• Custom cron expression
• Timezone selection

REST API Trigger

• Configure endpoint path
• Set authentication method
• Define request/response schema
• API rate limiting

Form Trigger

• Design form fields
• Enable CAPTCHA
• Set validation rules
• Configure submission handling

Step 3: Add Processing Nodes

Now add nodes to process the data from your trigger.

Adding Nodes to Canvas

1. Find the node type in the left palette
2. Drag and drop onto the canvas
3. Position it where you want
4. Release to place the node

Connecting Nodes

Connect nodes to define the data flow:

1. Click and hold on the output port (right side) of a node
2. Drag to the input port (left side) of another node
3. Release to create the connection
4. The connection line shows data flow direction

Node Organization

• Arrange nodes left-to-right to show workflow progression
• Group related nodes together
• Use the canvas zoom and pan controls
• Label nodes descriptively for team collaboration

Common Processing Patterns

Data Retrieval Pattern

Trigger → REST API Source → Transform → AI Gateway → Destination

Document Processing Pattern

Trigger → PDF Parser → Entity Extraction → MongoDB Destination

Conditional Pattern

Trigger → Conditional → [Branch A] → Destination A
                    → [Branch B] → Destination B

Step 4: Configure Nodes

Click any node to open its configuration panel on the right.

Basic Configuration

All nodes have these common settings:

Setting	Description
Node Name	Descriptive label displayed on canvas
Enabled/Disabled	Toggle node execution on/off
Notes	Documentation for team members

Node-Specific Settings

Each node type has unique configuration options:

Parameters

• Configure inputs, outputs, and filters
• Set data mapping from previous nodes
• Define transformation rules

Error Handling

• Retry logic and retry count
• Fallback values on failure
• Error notification settings

Execution Conditions

• When to execute this node
• Skip conditions
• Dependency requirements

Using Variables

Reference data from previous nodes using variable syntax:

// Reference output from previous node
{{ nodeName.outputField }}

// Access nested data
{{ apiSource.response.data.items[0].name }}

// Use in configuration
{
  "endpoint": "{{ trigger.endpoint }}",
  "payload": {{ dataTransform.result }}
}

Step 5: Test Your Workflow

Testing helps identify issues before deployment.

Running a Test

1. Click Test Run button in top toolbar
2. Provide test input data (for webhook/API triggers)
3. Watch execution in real-time on the canvas
4. Review output data and logs for each node

Real-Time Execution Visualization

During test runs:

• Nodes light up as they execute
• Progress indicators show current state
• Success/failure status displayed on each node
• Execution path highlighted

Reviewing Test Results

After execution completes:

Node Output

• Click any node to see its output data
• View formatted JSON/data structures
• Check data transformations

Execution Logs

• Console output for each node
• Error messages and stack traces
• Execution duration per node

Debug Mode

• Enable for verbose logging
• See intermediate processing steps
• Track variable values

Common Issues

• Missing Connections: Ensure all nodes are connected
• Invalid Configuration: Check for required fields
• Data Type Mismatches: Verify data formats match expected types
• API Credentials: Validate data source credentials

Iterative Testing

1. Fix any errors identified
2. Modify node configurations as needed
3. Test again
4. Repeat until workflow runs successfully

Step 6: Save Your Workflow

Before deploying, save your work:

1. Click Save button in toolbar
2. Enter workflow name
3. Add description (optional)
4. Click Save Workflow

Your workflow is now saved as a draft and ready for deployment.

Configuration Best Practices

Naming Conventions

• Use descriptive node names (e.g., "Parse Invoice PDF" not "Parser1")
• Follow team naming standards
• Include purpose in workflow name

Error Handling

• Add retry logic for external API calls
• Set fallback values for critical nodes
• Use conditional nodes for error routing

Documentation

• Add notes to complex nodes
• Document expected data formats
• Explain business logic in descriptions

Performance

• Minimize sequential dependencies
• Use parallel paths where possible
• Optimize data transformations
• Cache frequently accessed data

Next Steps

Now that you've created your workflow:

• Learn about all available nodes
• Configure workflow triggers
• Test your workflow thoroughly
• Deploy to production

Quick Start

For your first workflow, try a simple pattern: Webhook Trigger → AI Gateway → Webhook Destination. This creates a basic AI API endpoint.