This article provides a comprehensive guide to understanding and effectively using input nodes within the AgentRunner platform. Input nodes are essential components for defining and managing the data that flows into your agents. This guide will cover the various settings and options available for input nodes, including display names, privacy settings, input names, text fields, and history tracking. Whether you're a beginner or an experienced user, this guide will help you leverage input nodes to create more powerful and flexible agents.
Configuring Input Nodes
Input nodes in AgentRunner are highly customizable and allow you to define how data is received and processed within your agents. Understanding the different configuration options is crucial for creating efficient and reliable agents. This section will explore the key settings of input nodes, including display names, privacy controls, input names, and text fields. Each setting plays a unique role in how the input node functions within the agent, impacting both the design and execution of your agents.
Display Name
The display name of an input node is the label that appears in the agent editor. It serves as a visual identifier, making it easier to distinguish between different input nodes within a complex agent. The display name is purely for organizational purposes and does not affect the functionality of the node. You can choose any name that helps you quickly identify the node's purpose. There are no character limits or restrictions on the types of characters you can use for the display name, providing flexibility in how you label your nodes.
Private/Public Slider
The private/public slider determines whether the input node is visible to users when they run the agent. By default, this setting is set to "Public," meaning the input node is visible and editable by anyone running the agent. Switching the slider to "Private" completely hides the input from the Run Agent interface during runtime. This is particularly useful for locking inputs that should not be modified by end-users, ensuring that certain parameters remain constant. Making an input private does not prevent it from being seen in the logs.
Input Name
The input name is a crucial setting that defines the variable name you'll use to reference the input node's value in your code when calling the agent through an API. This name acts as a unique identifier for the input node, allowing you to pass data into the agent programmatically. It can only contain alphanumeric characters and the special characters "_", "+", and "-". Adhering to these naming conventions is essential for ensuring your agent functions correctly when called via API. The input name is case-sensitive, so ensure that you use the correct capitalization when referencing it in your code. It is important to note that you cannot use the same input name for multiple input nodes within the same agent. If you attempt to do so, AgentRunner will automatically append "_1" to the end of the second input name to ensure uniqueness.
Text Field
The text field is an optional field where you can specify a default or fixed value for the input node. This field also has an additional feature: to provide comments or instructions to other users about the purpose of the input node. The text field only accepts string data, but you can format the text as JSON if needed. If the text field is left blank, the input will be treated as empty. When the agent is executed, the user will need to manually fill in the value or provide it through the API; otherwise, the agent will not be able to run.
Monitoring Node History
AgentRunner provides a history feature for input nodes, allowing you to inspect the run history of each node directly from the node editor. This feature is invaluable for debugging, troubleshooting, and understanding how your agents are being executed over time. This section will cover how to access the node history, the type of information it provides, and any limitations associated with its use.
Accessing Node History
The history of an input node can be accessed directly from the node when opening it in the "History" panel. This panel provides a chronological list of all the times the node has been executed, along with relevant details about each run.
Information Included in Run History
The run history includes several key pieces of information for each execution of the node:
Timestamp: The date and time when the node was executed.
Run by: The username of the person who initiated the agent's execution.
Detailed log of the run: A comprehensive log of the node's execution, accessible by clicking a button associated with the specific run.
Limitations of the History Feature
While the history feature is a powerful tool, it does have certain limitations:
Access Restrictions: The history feature is available to those who can edit the agent, including owners and developers.
Exporting: The run history cannot be exported or downloaded for external analysis.
Configuration Changes: The history feature does not track changes to the input node's configuration (e.g., display name, private/public setting, text). It only tracks the node's run history.
Troubleshooting Input Nodes
Effective troubleshooting is essential for maintaining the health and performance of your AgentRunner agents. When working with input nodes, several common issues can arise that may prevent your agents from functioning as expected. This section will address these potential pitfalls and provide guidance on how to resolve them, ensuring your agents run smoothly and reliably.
Missing Input Values
One common issue is failing to provide a value for a public input when calling the agent through an API or running it manually. If an input's text field is left blank and no value is provided during runtime, the agent will be unable to run. Always ensure that all required public inputs have a value, either by setting a default value in the text field or by providing the value when executing the agent.
Incorrect Input Names
When calling an agent through an API, it's crucial to use the correct input name for each node. The input name is case-sensitive, so ensure you use the exact capitalization when referencing it in your code. Additionally, verify that the input name only contains alphanumeric characters and the special characters "_", "+", or "-". Any deviation from these rules can prevent the agent from running correctly.
