Workflow Builder

Creating a Workflow

Follow these steps to create and activate your first workflow:

1. Navigate to Workflows: From the main dashboard, click Workflows in the sidebar navigation.
2. Click "New Workflow": Click the + New button in the top-right corner of the Workflows page.
3. Name your workflow: Enter a descriptive name (e.g., "Order Confirmation via WhatsApp") and optionally add a description.
4. Set the trigger event: Select the event that will start your workflow from the trigger configuration panel. Use the category.action format.
5. Design on the canvas: Drag nodes from the palette, connect them with edges, and configure each node by clicking on it.
6. Activate: Toggle the workflow status to Active. The workflow will now listen for its trigger event and execute automatically.

Trigger Events

Trigger events are the entry points that start workflow executions. They follow a category.action naming convention. When an event is fired in your system, all active workflows listening for that event will execute.

The Visual Canvas

The workflow canvas is a React Flow-based editor (@xyflow/react v12) that provides a visual, drag-and-drop interface for building automation logic.

Drag Nodes from the Palette

The left-side node palette contains all available node types organized by category:

• Flow: Trigger, End, Branch, Delay, Loop, Transform
• Action: HTTP Request, Log Message, Send Notification
• Integration: Send Email, Send WhatsApp, WhatsApp Template, Create Shipment, Get Tracking
• AI: AI Prompt

Drag any node from the palette onto the canvas to add it to your workflow.

Connect Nodes with Edges

Click and drag from a node's output handle (bottom) to another node's input handle (top) to create an edge. Edges define the execution flow — data passes from one node to the next along these connections.

Configure Nodes

Click on any node to open its configuration panel on the right side. Each node type has specific configuration fields (see Node Reference for details).

Canvas Features

• Auto-save: Changes are saved automatically as you edit. No manual save required.
• Undo / Redo: Use Ctrl+Z / Ctrl+Shift+Z (or Cmd on macOS) to undo and redo changes.
• Zoom & Pan: Scroll to zoom, drag the background to pan.
• Multi-select: Hold Shift and drag to select multiple nodes at once.
• Minimap: A minimap in the bottom-right corner provides an overview of large workflows.

Variable Interpolation

Use the {{"{{}}"}} syntax to inject dynamic data into node configurations. Variables are resolved at runtime with data from the trigger payload and previous node outputs.

Trigger Data

{{trigger.field}}          — Access any field from the trigger event payload
{{trigger.customer.name}}  — Dot notation for nested objects
{{trigger.items[0].sku}}   — Array access with index

Previous Node Output

{{nodes.nodeId.field}}          — Access the output of a specific node by its ID
{{nodes.http_request_1.body}}   — e.g., HTTP response body
{{nodes.ai_prompt_1.response}}  — e.g., AI-generated text

Loop Context

{{loop.item}}    — The current item in the iteration
{{loop.index}}   — Zero-based index of the current iteration
{{loop.count}}   — Total number of items in the array

Transform Node & Pipes

The action.transform node allows you to reshape data between steps using pipe syntax for inline transformations.

Pipe Syntax

Pipes are applied using the | character after a variable reference. Multiple pipes can be chained:

{{trigger.name | upper}}         — Convert to UPPERCASE
{{trigger.name | lower}}         — Convert to lowercase
{{trigger.amount | default:0}}   — Use 0 if value is null/undefined
{{trigger.date | date:Y-m-d}}    — Format date as 2026-07-18
{{trigger.email | lower | trim}} — Chain multiple pipes

Transform Configuration

The transform node uses an assignments array where each entry defines a key-value mapping:

{
  "assignments": [
    { "key": "full_name", "value": "{{trigger.first_name | upper}} {{trigger.last_name | upper}}" },
    { "key": "total", "value": "{{trigger.amount | default:0}}" },
    { "key": "formatted_date", "value": "{{trigger.created_at | date:Y-m-d}}" }
  ]
}

Error Handling

Each node supports an on_failure configuration that determines what happens when the node encounters an error during execution.

Execution Statuses

Execution History

Every workflow execution is logged with full detail, allowing you to inspect, debug, and retry automations.

Viewing Executions

• Navigate to the Executions page from the main sidebar to view all executions across workflows.
• Alternatively, open a specific workflow and click the Executions tab to see only that workflow's runs.

Filtering

• By workflow: Select a specific workflow from the dropdown filter.
• By status: Filter by pending, running, success, partial, failed, or waiting.
• By date: Use the date range picker to narrow results to a specific timeframe.

Execution Detail View

Click any execution to open the step-by-step detail view:

• See each node that was executed in order
• Inspect the input data received by each node
• Inspect the output data produced by each node
• View error messages and stack traces for failed nodes
• See timing information (duration per node)

Retrying Failed Executions

Failed executions can be retried programmatically via the REST API:

POST /api/v1/executions/{execution_id}/retry
Authorization: Bearer {api_key}

Scheduling

Workflows can be triggered on a schedule using cron expressions, or by polling external APIs at configurable intervals.

Cron Expressions

Set a cron expression in the trigger configuration to run workflows on a schedule:

0 8 * * *      — Every day at 8:00 AM
0 9 * * 1      — Every Monday at 9:00 AM
*/15 * * * *   — Every 15 minutes
0 0 1 * *      — First day of every month at midnight
0 */2 * * *    — Every 2 hours

Polling Triggers

Polling triggers (trigger.polling) allow workflows to periodically check an external API for new data. Configure:

• URL: The external API endpoint to poll
• Interval: How often to poll (in seconds, minimum 60)
• Method: HTTP method (typically GET)
• Headers: Authentication headers for the external API

The polling engine deduplicates results to avoid processing the same data twice.