Skip to main content

Flow Call Return Node

The Flow Call Return node marks the exit point of a sub-flow and returns data back to the parent flow that initiated the call. It handles both text and JSON return values, allowing sub-flows to send processed results back to the calling parent flow. This node is the counterpart to the Flow Call Receiver Node in sub-flow architecture.

Flow Call Return node

Basic Usage​

Use the Flow Call Return node at the end of a sub-flow to send results back to the parent flow that called it.

Inputs​

The Flow Call Return node accepts the following inputs:

  • Returning Slot (Green handle): Receives the primary return value in text format from upstream nodes.
  • Returning Slot Json (Cyan handle): Receives complex return data in JSON format from upstream nodes.

Both inputs are optional, and the node will return the first resolved input value with text having priority over JSON.

Outputs​

The Flow Call Return node does not have outputs in the traditional sense. Instead, it sends data back to the parent flow:

  • Return Data: The text or JSON data received from input handles is sent back to the parent flow
  • Completion Status: A completion status is sent with the return data
  • Optional Jump Trigger: Can trigger the parent flow to jump to a specific node

Configuration​

Trigger Next Action​

A checkbox option that controls parent flow execution behavior:

  • Unchecked (Default): The parent flow continues executing normally after receiving the return value
  • Checked: The parent flow will jump to another specified node after receiving the return value

Use Case: Enable this when you want the sub-flow to determine which node the parent should execute next based on processing results.

Description: "Checking this box will cause the parent to jump to another node."

Example Workflows​

Employee Report Processing with JSON Return​

Scenario: Create a sub-flow that processes employee report text and returns structured JSON data to the parent chatbot flow.

Flow Call Return Example

Steps to Create the Sub-Flow:

  1. Start with a Flow Call Receiver node to accept the employee report text from the parent flow (chatbot).

  2. Add processing nodes for data extraction:

    i. Add a Text node with the employee report data (passed through from Flow Call Receiver)

    ii. Add a JSON Block node with the schema defining employee structure:

    {
      "type": "object",
      "properties": {
        "employees": [
          {
            "name": "string",
            "id": "string",
            "email": "string",
            "age": "number",
            "status": "string"
          }
        ]
      }
    }

    iii. Add a Text node with extraction instructions:

    Extract information for all employees mentioned in the report. Create an array of employee 
    objects and provide summary statistics.

    iv. Add a JSON Builder node:

    • Connect all inputs (report text, schema, instructions)
    • The JSON Builder processes and extracts structured employee data

  3. Add a Flow Call Return node at the end:

    i. Connect the JSON Builder output to Returning Slot Json:

    • Connect JSON Builder's JSON Output to the Returning Slot Json input handle (cyan connector)
    • This sends the structured JSON data back to the parent chatbot flow

    ii. Configure Trigger Next Action:

    • Leave unchecked for normal parent flow continuation
    • The chatbot will receive the JSON and present it to the user

  4. In the parent flow (Chatbot with Flow Call Processor):

    • User provides employee report text
    • Chatbot calls this sub-flow with the report
    • Sub-flow processes and extracts employee data
    • Flow Call Return sends JSON back to chatbot
    • Chatbot formats and displays the structured results to user

Preview:

The system will:

  1. Receive employee report text from chatbot via Flow Call Receiver
  2. Process through JSON Builder to extract structured data
  3. Generate JSON with employee information and summary statistics
  4. Send JSON back to chatbot via Flow Call Return (Returning Slot Json)
  5. Chatbot presents formatted employee data to user

Example Output JSON:

{
  "employees": [
    {
      "name": "John Smith",
      "id": "EMP001",
      "email": "john@company.com",
      "age": 32,
      "status": "active"
    }
  ],
  "summary": {
    "totalEmployees": 1,
    "averageAge": 32
  }
}

Result: Clean, modular sub-flow that processes unstructured text and returns structured JSON data to parent flows, enabling chatbots to present organized information to users.

  • Flow Call Receiver: Entry point for sub-flows, receives parameters from parent
  • Flow Call: Invokes sub-flows from parent flows and passes parameters
  • Flow Call Sender: Alternative term/implementation for returning data
  • Branching Node: Route to Flow Call Return based on conditions
  • Display Text: Show returned data in parent flow
  • Data Dump: Store returned data for debugging or logging
  • AI General Prompt: Process data before returning to parent
  • Execution Merger: Combine multiple processing paths before return

Parent Flow Handling with Trigger Next Action:

  • Sub-flow checks "Trigger Next Action"
  • Parent jumps to error handling node
  • Display friendly error message
  • Allow user to retry input