Test with mock data

Provide sample output for predecessor steps and run only the target step to test it in isolation without executing upstream steps.

Before you begin

Verify the target endpoint or model configuration before proceeding. REST API steps make real network calls and Task Agent steps incur real large language model (LLM) token costs during test runs.

Procedure

From a workflow on the canvas, click the step that you want to test.

The details panel opens.
From the Input pane, click Set mock data.

Select a step from the dropdown menu to open the mock data editor. The steps that display in the dropdown list depend on which steps your workflow includes.
In the JSON editor, provide mock output data for the predecessor steps.
This data represents the output that each predecessor would produce in a real run.

Note:

If the step has output from a previous run, the editor pre-populates with that data instead of an empty template. You can edit the pre-populated values before pinning them.

For example, if the target step references ${nodes.fetch_data.output.hosts}, provide mock output that includes that structure:
```
{
  "status": "completed",
  "stdout_json": {
    "hosts": ["web-01", "web-02"]
  }
}
```
Optional: Click Pin data to preserve the mock data.

You can use pinned data when running this step in combination with other steps.

Note:

You can unpin data individually or all at the same time.
Repeat this process if you want to test with mock data on other steps at this time.
Click Run step.

The Set mock data dialog opens and shows the mock data you set earlier. Click Run from this dialog.

The system executes the target step and uses the mock data as input from its predecessors. Credential resolution and template expression evaluation run as they do in a production workflow. Execution stops after the target step completes.
Click Save.

The workflow displays in the canvas. You can see the status and which steps the test run skipped. After the test run, steps that received mock data display a Mock data pinned badge on the workflow canvas.
Click Back to editor to open the three-pane view and see the output of the test run in the Output pane.
From the Parameters pane, review the execution content.

Results

Confirm that the target step status shows as completed or failed.
Verify that the output data matches the expected response for the step type.
Check the test run record in the run history by filtering for test mode runs.