Connections define how data flows from one node to another. Understanding connection types and variable mapping is essential for building effective workflows.
Creating Connections
To connect two nodes:
- 1
Click and drag from an output handle Output handles appear on the right side of each node
- 2
Release on an input handle Input handles appear on the left side of nodes
- 3
Configure variable mapping The mapping editor opens automatically after connection
Connection Types
Direct Connections
Simple one-to-one data flow from one node to another. The entire output of the source node is passed to the input of the target node.
Mapped Connections
More granular control where you specify exactly which fields from the source should map to which inputs on the target. This is the most common connection type.
Multiple Connections
A single output can connect to multiple target nodes, allowing you to branch your workflow and process data in parallel paths.
Variable Mapping Editor
When you create or edit a connection, the mapping editor allows you to:
- Select source variables: Choose from
out,_input, or specific fields likeout.fieldName - Map to target inputs: Choose which input on the target node receives the data
- Preview data structure: See the shape of data flowing through the connection
- Multiple mappings: Create multiple mappings for complex data transformations
Common Connection Patterns
Sequential Processing
Connect nodes in a linear chain where each node processes the output of the previous node. Example: Dataset Source → Prompt → Dataset Sink
Parallel Processing
Split data to multiple processing paths. Use an Array Splitter to process items individually, then Array Flatten to recombine results.
Data Enrichment
Pass original input through while adding processed results. Use _input to access the original data alongside out for the processed output.
Deleting Connections
To remove a connection:
- Click on the connection line to select it
- Press Delete or Backspace
- Or right-click the connection and select "Delete"