Data structure

In Agentic WorkFlow, every node takes some input data and produces output data.

If you understand the “shape” of that data, it becomes much easier to:

Connect nodes correctly
Know what the next node will receive
Debug when something looks wrong

How data flows between nodes

Every node:

Receives data
Does its job
Outputs new data for the next node

Most nodes output a list of items.

Each item is a JSON object (think: one “record” with fields).

[
  {
    "name": "Alice",
    "email": "alice@example.com"
  },
  {
    "name": "Bob",
    "email": "bob@example.com"
  }
]

If there are multiple items, the next node runs once per item.

Multiple inputs and data merging

Some nodes can receive data from more than one input.

When that happens, Agentic WorkFlow combines items into all possible combinations.

In plain terms: if one side has 2 items and the other side has 3 items, you’ll get 6 combined items.

flowchart TB
    B["Node 1"] L_B_D_0@--> D["Node 3"]
    C["Node 2"] L_C_D_0@--> D
    n1[" "] L_n1_B_0@--> B & C

    B@{ shape: rounded}
    D@{ shape: rounded}
    C@{ shape: rounded}
    n1@{ icon: "fa:circle-play", pos: "b"}

    L_B_D_0@{ animation: slow } 
    L_C_D_0@{ animation: slow } 
    L_n1_B_0@{ animation: slow } 
    L_n1_C_0@{ animation: slow }

If Node 1 outputs n items and Node 2 outputs m items, then Node 3 receives n × m combined items.

Example

Node	Output
Node 1	`[{"id":1}, {"id":2}]`
Node 2	`[{"color":"red"}, {"color":"blue"}]`
Node 3 (combined input)	`[{"Node 1": {"id":1}, "Node 2": {"color":"red"}}, {"Node 1": {"id":1}, "Node 2": {"color":"blue"}}, {"Node 1": {"id":2}, "Node 2": {"color":"red"}}, {"Node 1": {"id":2}, "Node 2": {"color":"blue"}}]`

This is useful when you want to “pair up” data from two sources.

Tips for building data-driven workflows

Check your data at any step with the node output/preview panel in the workflow editor.
If you need to reshape data, use transformation nodes (for example “Edit Fields”).
When combining multiple branches, see the flow logic pages.

Data structure

How data flows between nodes

Multiple inputs and data merging

Example

Tips for building data-driven workflows

Related topics