Data Flow & Conditions

What you'll be able to do after this module

• Write template variable expressions to reference any upstream node's output
• Understand how the variable system resolves data at execution time
• Build conditional branches that route execution based on runtime values
• Apply the three core data passing patterns: linear chain, fan-out, and conditional routing

Template Variables

Template variables are how nodes communicate. When you type in an action's config field, you can reference the output of any upstream node using this syntax:

{{@nodeId:NodeLabel.field}}

How to insert variables

1. Click an action node to open the config panel
2. Place your cursor in a template-enabled field (text input or textarea with the {{ }} icon)
3. Type {{ — a dropdown appears showing all upstream nodes and their available output fields
4. Select a node and field to insert the full reference

Examples

{{@trigger:Webhook.body.customer_id}}     → access webhook payload field
{{@node1:Scrape URL.markdown}}             → scraped page content
{{@node2:Generate Text.text}}              → AI-generated text
{{@node3:Get Customer.email}}              → nested field (auto-unwrapped from data)

Data resolution at runtime

When a step executes, the engine:

1. Looks up the stored output for the referenced nodeId
2. Automatically unwraps the standard { success, data, error } envelope — so .text actually accesses data.text
3. Substitutes the resolved value into the config field before the action runs

If a referenced node hasn't executed yet (e.g., it's in a different branch), the variable resolves to undefined.

Data Passing Patterns

There are three fundamental patterns for wiring data through a workflow:

Pattern 1: Linear Chain

Each step reads from the step before it.

• Generate Text reads {{@A:Scrape URL.markdown}}
• Send Email reads {{@B:Generate Text.text}}

This is the most common pattern — a sequential pipeline.

Pattern 2: Fan-Out

Multiple steps read from the same upstream node.

Here, the classifier output is consumed by three different actions. All three run after Classify Ticket completes.

Pattern 3: Conditional Routing

A Condition node evaluates an expression and routes to one of two branches.

The Condition checks whether the score meets a threshold. Only the matching branch executes.

The Condition Node

The Condition is a system action (no integration needed) that evaluates a boolean expression and routes execution to a true or false branch.

How to add a Condition

1. Click + to add a new node
2. Select Condition from System Actions
3. Connect two outgoing edges from the Condition node — one for the true path, one for the false path

Writing Expressions

The expression field accepts template variables, literals, and operators:

Comparison operators:

Logical operators:

Value types:

Expression examples

// Simple comparison
{{@node1:Get Customer.status}} === "active"

// Numeric threshold
{{@node2:Score Lead.score}} > 80

// Compound logic
{{@node1:Get Customer.status}} === "active" && {{@node2:Score Lead.score}} > 50

// Negation
!{{@node3:Guard.detected}}

Branching behavior

• The true edge connects to the actions that run when the expression evaluates to true
• The false edge connects to the alternate path
• Each branch can have any number of downstream steps
• Only the matching branch executes — the other is skipped entirely

Practical Exercise: Priority Router

Build a workflow that classifies incoming support tickets and routes them based on priority.

{
  "subject": "string",
  "body": "string",
  "customer_email": "string"
}

{
  "subject": "Cannot access dashboard",
  "body": "I'm locked out of my account and have a demo in 30 minutes",
  "customer_email": "[email protected]"
}

What you now understand

Up next: Triggers & Webhooks — make your workflows callable from the outside world.