Skip to content
Rival Docs

Running & Testing

Overview

While you build, you’ll want to run the workflow to see it work. Rival gives you three ways to do that:

  • Run the whole workflow and watch it live.
  • Test one step at a time to check its input and output.
  • Review past runs in the run history.

Running the whole workflow

Click the blue Run button in the toolbar. The workflow executes end to end, and the right-hand Run log panel streams live progress:

  • A status chip - running, succeeded, or failed.
  • A card for each step as it runs, showing its output.

Steps run in order, following the connections you drew. Branch steps send the flow down the matching path, and a step that fails stops the run and marks it failed.

Testing one step at a time

Open a step’s detail view to test it in isolation. It has three panes - INPUT on the left, Parameters in the middle, and OUTPUT on the right - with two buttons:

  • Execute previous nodes - runs the steps before this one and fills in its INPUT, so you can see exactly what this step receives.
  • Execute step - runs through this step and fills in its OUTPUT.

This is the fastest way to debug a single step without running the whole flow. Configure the step, execute it, and check the OUTPUT pane. If a step emits an interactive output (like an agent’s artifact), it renders right there in the OUTPUT pane.

Run history

Click History in the toolbar to open the Run history panel (you can also open the builder with the history panel already showing).

Each row shows:

  • A status dot - succeeded or failed.
  • When it ran (relative time).
  • The run’s cost.

Click a row to reopen that run and replay its per-step results. Use Refresh to pull in the latest. If you haven’t run the workflow yet, the panel reads “No runs yet. Hit Run to create one.”

Next steps