Skip to main content
Use the Play you selected and the CSV you prepared. Your agent starts the run, keeps its ID, and follows it through completion. Review the sample and available usage estimate before starting; a run can spend Deepline credits.

Start in the web app

  1. Open your selected Play from Tools > Prebuilt plays, or find a saved Play in All plays. Check that you are in the intended workspace.
  2. Open the start control above the Play’s steps. It is labeled Run here, or Start when a trigger is configured. Choose Run here from its menu.
  3. Supply the required fields or attach your sample CSV. Check the inputs, then select Run to start the work.

Review and export the result

Inspect the selected run’s status, output, and errors. Return to it through the Play’s run history or Recent runs. A run started by your agent appears here too; give your agent the run ID to inspect a run you started in the app. Open the run record to inspect its output. When the run writes a dataset, use its Download … CSV control to download that run’s rows, or ask your agent to export the saved run’s dataset. Keep the run ID with the exported file. A completed run can still contain no matches or unresolved rows.
  • Check a few representative rows before running the full file.
  • Review results against your acceptance rule; completion alone does not make every row usable.
  • If waiting stops or a response is lost, inspect the saved run before starting another. Reading or exporting saved results does not rerun the Play.
Next, use the saved output to Visualize a saved result. Agent guide: CLI and SDK instructions →