Data Flow & Context

The Execution Context

Every node receives an ExecutionContext instance that contains everything it needs to operate. Understanding the context is essential for building correct workflows.

class ExecutionContext
{
    public array $input;       // Output from the immediately preceding node
    public array $config;      // This node's configuration
    public array $sourceData;  // Original data from the trigger node
    public Execution $execution;
    public Node $node;

    // Methods
    public function getConfig(string $key, mixed $default = null): mixed;
    public function getFullContext(): array;
    public function setVar(string $key, mixed $value): void;
    public function getVar(string $key, mixed $default = null): mixed;
    public function getNodeResult(string $nodeId): array;
}

Data Propagation Model

Sequential $input

Each node receives the output of its direct predecessor as $context->input:

Trigger → [output: {timestamp: ...}]
              ↓ input
Data Model → [output: {data: [{id:1,...}, {id:2,...}]}]
              ↓ input
Filter → [output: {data: [{id:1,...}]}]  (filtered)
              ↓ input
Send Mail

Source Data $sourceData

The trigger's original payload is preserved throughout the entire workflow in $context->sourceData. This allows any node to access the original event data even after it has been transformed by intermediate nodes.

Global Variables

The Set Variable node writes to $context->vars. Any downstream node can read these values using {{variables.key_name}} in templates or $context->getVar('key_name') in PHP Code nodes.

Variables are persisted in the context JSON column of the Execution record, so they survive across async job dispatches.

Full Context Object

$context->getFullContext() returns a merged array combining:

[
    // Trigger payload (top-level)
    'event' => '...',
    'user_id' => 42,

    // Input from previous node
    'input' => [...],

    // Global variables
    'variables' => [
        'my_var' => 'value',
    ],

    // Results from named nodes
    'nodes' => [
        'node_uuid' => ['output' => [...]],
    ],

    // Current iteration item (inside For Each)
    'item' => [...],
]

This merged object is what {{tag}} resolution runs against.

Data Normalization

Voodflow normalises data passing through nodes:

• Arrays are passed as PHP array (not objects)
• Eloquent models are converted to arrays via ->toArray() before entering the pipeline
• Null values are preserved
• Non-serializable objects are logged as warnings and stripped

Context Size Management

The context object grows as nodes add data. For large datasets, avoid accumulating all records in the context. Instead:

• Use the Aggregate node to reduce large arrays into summaries
• Use the Transform node to project only the fields you need downstream
• Set VOODFLOW_CONTEXT_SIZE_WARNING_KB to get warned when context grows large