Skip to content
synthreo.ai

Input Data - Synthreo Builder

Input Data node for Builder - define the entry point for your AI agent workflow, mapping incoming request fields and parameters to workflow variables for downstream nodes.


The Input Data node is the starting point for Builder workflows. It acts as the entry door where data first arrives in your workflow. This can include inputs from user interactions, scheduled triggers, or file uploads.

This node is essential for any workflow that handles external input. It defines the shape of data entering the workflow and exposes that data to all downstream nodes as named variables.


  • None. This node is always the first node in a conversational or triggered workflow and has no upstream connection.

The node produces a JSON object containing the incoming data. By default, the object carries the key userSays holding the user’s input text. The output variable name is controlled by the Output Column Name parameter.

Downstream nodes reference the output of this node using the {{Out.outColumnName}} syntax - for example, {{Out.userSays}} with the default name. In many prompt fields you can also use just {{userSays}} when the variable name itself is unambiguous.


NameTypeRequiredDefaultDescription
Input is JSON (inputIsJson)Toggle (On/Off)OptionalOffEnable when the agent receives a structured JSON payload (e.g., from a scheduled trigger or another agent) rather than a plain text message. When enabled, the incoming payload is parsed as a JSON object and each top-level key becomes accessible as a separate variable.
Contains File (containsFile)Toggle (On/Off)OptionalOffEnable when the input may include an uploaded file. Activates the Base64 Value Expression field.
Base64 Value Expression (base64ValueGetter)TextRequired if Contains File is onbase64fileThe key name under which the Base64-encoded file data arrives in the incoming payload. This value is exposed downstream as {{base64file}} by default. Change it if your payload uses a different key name for the file data.
Save File (saveFile)Toggle (On/Off)OptionalOffWhen enabled, the uploaded file is persisted to system storage. Useful for compliance requirements or when the file needs to be accessed by later workflow runs.
Folder Path (folderPath)TextRequired if Save File is onEmptyThe storage path where uploaded files are saved. Supports dynamic variables, e.g., /invoices/{{currentYear}}.
Output Column Name (outColumnName)TextOptionaluserSaysThe variable name that downstream nodes use to reference this node’s output. With the default value of userSays, downstream nodes reference the input as {{Out.userSays}} or {{userSays}}. Change this to a more descriptive name for workflows that process something other than a user chat message.

After the Input Data node runs, all downstream nodes have access to its output variable. The variable name is whatever you set in Output Column Name (default: userSays).

In LLM node prompts:

Place {{userSays}} anywhere in the System Message or Prompt Editor to inject the user’s input into the model’s context.

You are a helpful assistant. The user has asked: {{userSays}}

In downstream node fields that accept expressions:

Most node fields that accept variable references use the {{Out.variableName}} form:

{{Out.userSays}}

When Input is JSON is enabled:

If the incoming payload is a JSON object like {"orderId": "12345", "customer": "Acme Corp"}, each key in that object becomes directly accessible in downstream nodes. You can reference {{Out.userSays.orderId}} or use dot-notation to access nested properties.


  • Conversational agent: Leave all settings at their defaults. userSays is populated from the chat message and referenced as {{userSays}} in a downstream LLM prompt.
  • Scheduled agent: Enable Input is JSON so the structured payload from the scheduler (which contains trigger metadata and any configured parameters) is parsed correctly. Reference individual fields from the payload in downstream nodes.
  • File processing workflow: Enable Contains File, then reference {{base64file}} in a downstream file processing node such as the File to Text or OCR node. Enable Save File and set Folder Path if the original file needs to be retained.
  • Agent-to-agent call: When one Builder agent calls another, enable Input is JSON on the receiving agent so the structured data from the calling agent is parsed into accessible variables.

  • Always use this node as the first step in workflows that handle external input.
  • Enable Input is JSON for any agent that runs on a schedule or is called by another agent. Scheduled triggers send a JSON payload that will be silently ignored without this setting.
  • Choose a meaningful Output Column Name when the workflow handles something other than a user chat message (for example, incomingOrder for an order processing workflow). This makes variable references in downstream nodes easier to read and maintain.
  • Use downstream nodes (e.g., LLM, Custom Script, Transform) to reshape or validate the data before further processing rather than expecting the Input Data node to do filtering or transformation.
  • When Save File is enabled, use dynamic expressions in Folder Path (e.g., including a date or user ID) to keep saved files organized and avoid overwrites.