Skip to main content

ThreoBuilder Debugger

The Debugger in ThreoBuilder is a troubleshooting tool that allows you to inspect and understand what happened during a workflow run.

It provides visibility into each node’s triggered actions, the data passed between them, and additional execution details — helping you pinpoint issues or verify behavior.

🎯 Purpose​

  • Review the latest values processed by each node
  • Inspect the input data received by nodes
  • See detailed execution logs and messages for each triggered node

📝 How to Access the Debugger​

  1. Open ThreoBuilder and load the workflow you want to inspect
  2. Click the Test button in the top toolbar to manually execute the workflow
    • The Debugger tab is only available after a manual execution
  3. In the Test Results panel, click the Debugger tab
  4. The Debugger view will display all nodes that were triggered during the run

alt text

📊 Understanding the Debugger View​

The Debugger contains two main tables:

1️⃣ Node Execution Table​

Shows all nodes that were triggered in the workflow.

ColumnDescription
Node NameThe name of the triggered node
ActionsClick the ℹ️ icon to view the input data that the node received

Example:

  • Input Data – Shows the initial data passed into the workflow
  • Azure OpenAI – Displays the prompt or parameters sent to the model
  • Output Data – Shows the final processed data before workflow completion

2️⃣ Node Details Table​

Displays additional execution messages for each node.

ColumnDescription
Stack NameThe node name where the message originated
MessageDetails about the node’s execution
ActionsClick the ℹ️ icon to view more context or raw data

Example:

  • Azure OpenAI – RAG selection items: Shows the records chosen by the Retrieval-Augmented Generation process based on its understanding
  • Azure OpenAI – Node execution completed: Confirms successful execution

💡 Example Usage​

  • Verifying Data Flow – Check exactly what data each node received during execution
  • Inspecting AI Behavior – Review RAG selection items to understand why certain records were chosen
  • Debugging Errors – Identify which node failed and inspect its input to find the cause

📘 Best Practices​

  • Always run the workflow manually via Test before opening the Debugger
  • Use the ℹ️ icon to drill down into node inputs and execution details
  • Review both tables to get a complete picture of the workflow’s behavior
  • When testing AI nodes, pay attention to RAG selection details for accuracy checks