Connections

A connection (also called an edge) links one node's output handle to the next node's input handle. The engine follows these connections in graph order to determine which nodes run next and what data each node receives.

How edges work

Each edge has a source node and a target node. It optionally specifies a source handle (the named output port on the source node) and a target handle (the named input port on the target node). When the source node finishes, its output value is injected into the target node's named input before that node runs.

To draw a connection on the canvas: drag from the dot on the right edge of a node (output handle) to the dot on the left edge of another node (input handle).

Branch edges

Some nodes have multiple output handles that represent alternative paths:

• Conditional (CONDITIONAL) — exposes true and false output handles. Only the branch that matches the evaluated condition continues.
• A/B Test (AB_TEST) — exposes a and b output handles. One branch is selected per run.
• String Condition and Number Condition — also expose true/false branches.

You wire each branch handle to a different downstream node. When the engine evaluates the node, exactly one branch proceeds; the other is skipped.

Data passing between nodes

Each node's output payload is available to all nodes further downstream in the graph. In a node's configuration panel you can reference upstream values using the {{variableName}} syntax — for example, {{data.phone}} to use a field from the HTTP Trigger's request body.

The canvas shows which variables are available at each node based on what its upstream nodes produce. See Variables and data for the full reference on writing expressions and mapping values.

Related

• Nodes — the input and output handle names for every node type
• Workflows — how nodes and connections form a DAG
• Variables and data — referencing upstream values in node configuration