Ez az agent futások naplója. Itt látszik, milyen inputtal futott egy agent, mit válaszolt, mennyi ideig tartott, milyen modellt használt és volt-e hiba. Debugoláshoz és költségkövetéshez fontos.
Trace agent runs, retrieved chunks, tool calls, latency, errors, and exportable execution evidence.
Inspect every run as a supportable event: prompt messages, retrieved chunks, tool calls, latency, errors, and exportable JSON.
Execution rows in current filter.
0 succeeded in this view.
Calculated from loaded rows.
Live executions still in progress.
Exportable JSON evidence remains available.
Load a workspace to inspect execution traces, filter by status, and open detail evidence.
Select an execution to inspect its full trace.
Choose a run from the left panel to inspect messages, retrieved chunks, tool calls, latency, cost placeholders, and exportable JSON.