Run agent (synchronous)
Run the agent with a query and return a JSON response.
Body
Response
Successful Response
Response model for synchronous agent run.
Name of the agent
Session ID for this conversation
Approvals a paused run is waiting for
The budget a waiting run ran out of, and what it needs
Optional metrics for this run
Agent's response (none while the run awaits approval)
Runtime run ID for this run
When the run ended in error: the calls it already made that may have changed something (tool name, call id, outcome success or unknown). Resume the run; a new run can repeat them.
Runtime outcome: success, error, cancelled, awaiting_approval, or awaiting_budget
Why execution terminated
Telemetry trace ID for this run