Skip to main content
Executions

Runs

See and monitor every run of the agent, separated by environment.

Runs

The Runs page lists every run of the agent, so you can watch the history live and inspect each one.

Reach it through the Runs icon in the editor's side navigation.

Environments

Runs are split across three tabs:

TabWhat lands there
ProductionRuns started by real users through the chat, the integrations (Teams, WhatsApp and so on) or the API, in production
DevelopmentRuns from the editor's playground, or from development
TestsRuns generated by the test sets

The runs table

ColumnWhat it is
ModelThe language model used (Roberty LLM, GPT-4o, Claude, Gemini, LLaMA)
ModeHow the run was started: Chat, Prompt, Form or API
TemperatureThe temperature configured when the run started
StartedHow long ago it began (e.g. "2 hours ago")
StatusWhere the run currently stands
DurationThe run's total time (once it has finished)

Run statuses

StatusColourWhat it means
FinishedGreenThe run completed successfully
FailedRedSomething went wrong during the run
RunningBlue (blinking)The run is in progress
PendingYellowQueued, waiting to start

The list refreshes itself while any run is in progress.

Opening the details

Click any row to open the run details panel, with everything about the inputs, outputs, trace and logs.