Skip to main content

7. Testing and Running

1. Testing a Single Node​

To test a node individually:

  1. Select the node.
  2. Click the Test button on the node or in the right panel.
  3. The node runs using the data received up to that point.
  4. You can view the results inside the node or in the right panel.

This method is especially useful for debugging:

  • HTTP Call nodes
  • OpenAI / Anthropic / SkyStudio API Call nodes
  • Output Text nodes
  • Query nodes

2. Running the Entire Workflow​

To test the entire flow as a complete scenario:

  1. Click the Run button in the top menu.
  2. The test panel opens.
  3. Fill in the inputs required by the starting nodes, such as Text Input, File Input, or Webhook.
  4. Click the Submit button.
  5. The workflow runs from start to finish, and you can monitor the results from the panel.

3. Debugging Tips​

  • 🔍 Proceed Step by Step: In complex flows, test each node separately to narrow down the source of the error.
  • Check JSON Outputs: Review the field names in the JSON outputs of nodes such as HTTP, LLM, and Parser. An incorrect field name is often the cause of an error.
  • Follow the Connection Flow: Especially in If-Else nodes, check whether the TRUE and FALSE branches work as expected.