You never have to run a flow blind. Every stage will show you its rows before you commit to a full run.
Preview a single stage
Double-click a stage to open it. The "Configure" tab holds its settings; the "Preview" tab shows a "Sample rows" panel with its data.
If you see "No sample rows available yet for this stage.", the stage hasn't got anything to show yet — some stages fill their preview in only after the flow has run once.
Compare in and out
Select a stage and the dock along the bottom of the canvas shows its data from the last run. This is the fastest way to check whether a filter or a join did what you meant:
- Switch between "Data", "Map", and "Logs" for rows, geometry, and log lines.
- Flip the "In" / "Out" toggle to compare what entered the stage with what left it.
- "CSV" downloads the rows you're looking at.
- Right-click a cell for "Copy cell", "Copy row", and "Copy column name".
Watch for a "Flow changed since this run" badge — it means you've edited the canvas since these rows were produced, so re-run before trusting them.
Run the flow
"Run flow" in the top bar runs everything; "Cancel run" stops it. A status chip tracks progress while it works.
To iterate on the middle of a long flow, right-click a stage and choose "Run to this stage" — "Runs this stage and everything before it". It saves you re-running work you've already validated.
If the run won't start
Stages that still need attention are flagged on the canvas as you edit, and a run won't start while any remain: "Run blocked: N stages need attention before execution." Open each flagged stage, finish its configuration, and run again.