Skip to main content
The Output node defines the exit points for your workflow, allowing you to pass data out to parent workflows, webhook responses, subflows, and external systems. It’s essential for creating reusable workflows and enabling agents to access your flow results. Output node interface

​
Quick Start

1

Add the Output node to your workflow

Drag the Output node from the node library into your canvas at the end of your workflow
2

Connect your data

Drag the output badge from a previous node into the Output field
3

Name your output

Give your output a descriptive name (e.g., “summary”, “processed_data”, “email_list”)
4

Set the output type

Choose the appropriate data type: Text, List, or Any

​
Node Configuration

​
Required Fields

The value or data you want to pass out of the workflow. Connect this to the output of any previous node in your flow by dragging its output badge into this field.
A descriptive name to identify this output. This name is used when:
  • Accessing the output in parent workflows
  • Retrieving data via webhook responses
  • Using the flow as a subflow in other workflows
  • Allowing agents to view the results when the flow is used as a tool
Default: “output”

​
Optional Fields

Sets the expected data format for your output:
TypeUse CaseExample
TextSingle values like strings or numbersA summary, email address, or processed text
ListArrays of valuesList of URLs, email addresses, or extracted items
AnyMixed or unknown data typesAPI responses, complex objects, or testing

​
Multiple Outputs

You can configure multiple outputs within a single Output node by clicking the + Add outputs button. This is useful when your workflow produces several distinct results that need to be accessed separately.

​
When to Use the Output Node

Subflows

Required for reusable workflowsWhen creating a subflow, the Output node defines what data becomes available to the parent workflow. Without it, the parent flow cannot access any results.

Webhooks

Return data to external systemsWhen your workflow is triggered via webhook, the Output node’s data is returned in the API response, making it accessible to the calling system.

Agent Tools

Enable agents to view resultsWhen using a flow as a tool in an agent, the Output node is required for the agent to see and use the results of the workflow execution.

API Responses

Programmatic accessWhen running workflows via the Gumloop API, outputs are returned in the get_pl_run response, allowing programmatic access to results.

​
Using Flows as Agent Tools

If you’re using a flow as a tool in an agent, you must include an Output node for the agent to view the results of the workflow execution.
When an agent invokes a flow as a tool, it needs to receive the results to continue its reasoning and decision-making. Without an Output node, the agent cannot see what the workflow produced, making the tool effectively useless. Output node connected to Ask AI for agent tool usage

​
Example: Creating an Agent-Compatible Flow

Consider a flow that uses Ask AI to analyze text. To make this flow usable as an agent tool:
1

Build your workflow logic

Add your processing nodes (e.g., Ask AI, data extraction, API calls)
2

Add an Output node at the end

Connect the final result to an Output node
3

Name the output descriptively

Use a clear name like “analysis_result” or “summary” so the agent understands what it’s receiving
4

Save and add to your agent

The agent can now invoke this flow and receive the output to use in its reasoning
The output name you choose will be visible to the agent, so use descriptive names that help the agent understand what data it’s receiving.

​
Working with Subflows

The Output node is essential for creating modular, reusable workflows through subflows.

​
Passing Data to Parent Workflows

When you use a flow as a subflow, all outputs defined in the Output node become available in the parent workflow: 
Subflow outputs: "customer_name", "customer_email", "order_total"
Parent flow: Can access all three outputs when using the subflow node

​
Chaining Subflows

Create complex data processing pipelines by passing outputs from one subflow to another: 
Subflow A (Data Extraction) -> Output: "extracted_data"
    |
    v
Subflow B (Data Processing) -> Input: receives "extracted_data"
    |
    v
Output: "processed_result"
Name your outputs clearly and consistently across subflows to make your workflows easier to understand and maintain.

​
Output Types Explained

  • Text
  • List
  • Any
Use for single valuesChoose Text when outputting:
  • A single piece of text or string
  • A number or calculated value
  • A summary or processed result
  • Any single, non-list value
Example: "The analysis shows a 15% increase in engagement."

​
Common Use Cases

Extract and output email addresses from a document:
Document Input -> Extract Data -> Output (name: "email_list", type: List)
The parent workflow or webhook can then use this list for further processing.
Summarize content and return the result:
Content Input -> Ask AI (summarize) -> Output (name: "summary", type: Text)
Perfect for creating reusable summarization subflows.
Transform data and output multiple results:
Raw Data -> Process -> Output (names: "cleaned_data", "error_count", "processing_time")
Use multiple outputs to provide comprehensive results.
Create a research flow that an agent can use:
Search Query Input -> Web Search -> Summarize Results -> Output (name: "research_findings")
The agent receives the research findings and can use them in its response.

​
Important Considerations

  • Use descriptive, lowercase names with underscores (e.g., customer_email, processed_data)
  • Avoid generic names like “output” or “result” when possible
  • Keep names consistent across related workflows
  • Consider how the name will appear to agents or in API responses
  • Match the output type to your actual data to avoid unexpected behavior
  • Use List type when working with Loop Mode results
  • Use Any type sparingly, as it provides less clarity to downstream consumers
  • Add multiple outputs when your workflow produces distinct results
  • Each output can have its own name and type
  • All outputs are available simultaneously to parent workflows and API responses
The Output node is your workflow’s gateway for sharing results with the outside world. Whether you’re building reusable subflows, creating webhook-triggered automations, or enabling agents to use your flows as tools, properly configured outputs ensure your data flows seamlessly to wherever it needs to go.