Workspace / RunsConnecting

OVERVIEW

Runs

Your research, from first instruction to final result.

Active runs Loading capacity…
Completed Finished with all required evaluations
Needs attention Failed or incomplete runs
Total runs Across your workspace

Run history

Open a run to follow its activity and inspect results.

Updates automatically
Run Status Builder Activity Elapsed Created Open run

Loading your runs…

Every run keeps a complete trajectory, its specification, and collected artifacts.

START SOMETHING NEW

New run

Give the agent a clear research spec. You can follow its work and send guidance while it runs.

Evaluate on

Every selected model must receive a score or a classified outcome for this run to complete.