Agent Runs, Pro
Every run of your AI agents, a page at a time from your server, with its task, status, steps, duration and cost, and the runs waiting for a person to approve what they ask.
The list behind an agent console: the runs that need someone first, then every run, to find the ones that failed, ran long or cost too much.
Runs
Three refunds waiting for a person, longest first, above every run you filter by status, agent or customer, with each run's outcome or error under its status.
Waiting for a person
Longest wait first. Each run holds its refund and reply until someone approves or denies it.
All runs
Every agent run in Acme.
| Task | Status | Agent | Customer | ||||
|---|---|---|---|---|---|---|---|
| Loading | |||||||
Agent Run Detail
One run of an AI agent opened in full, with what it did, what it cost, the decision it waits on, and every step as a timeline or a trace.
Agent Session Timeline
One run of an AI agent, live, with what it is doing now, its plan, every model call and tool call, the approvals it waits for, its reply, and what it cost.