Read a run and its live trace
Follow an agent step by step while it works, understand each status and open the details of any tool call.
Every run keeps a complete record. You can watch it as it happens or read it later.
Open a run
- From Run agent, you are taken straight to the run.
- On the Agents page, the Recent runs list shows the latest runs; choose one.
- On an agent's page, the Recent runs panel lists its last runs.
The page title reads "Agent name: run" with the trigger and start time below. A Live label appears while the run is in progress.
The header
- A status badge: Queued, Running, Waiting for your approval, Done, Failed or Cancelled.
- A summary line: number of steps, tokens and cost.
- Open agent returns to the agent. Cancel run appears while the run is active.
- Input shows what the run was given, if anything.
The steps
Below that is a numbered list. Each step has a title and, where it applies, the tokens and cost it used:
- Thinking: the agent's reasoning.
- Calls : the agent used a tool. Choose Show details to see exactly what it sent.
- returned (green) or failed (red): what came back. Show details reveals the full result.
- Needs your approval: the run paused for you.
- Answer: the agent's reply.
- Stopped with an error: something went wrong.
When the run finishes, a Result box shows the final output. If it failed, a red box shows the reason in plain words.
Resuming after a pause
If a run is Waiting for your approval, the approval card appears at the top of the page. Press Approve or Deny and the run continues from the same point. See The approvals inbox.
Good to know
- If your connection drops, the page falls back to refreshing every few seconds.
- A run that fails because of a temporary problem is retried automatically a few times before it gives up. Unexpected errors show a reference starting with ERR- that you can quote to support.
- Fetched pages and tool results are shown in the trace, so avoid sending secrets through tools.
- The trace is your best debugging tool: if the agent did something odd, the cause is almost always in the steps above its answer.
More guides in PGB Agents
What a PGB Agent is and what it can do
An agent is a saved job with instructions, a list of allowed tools, a budget and a record of every step it takes.
Create an agent from scratch or from the Marketplace
Make a blank agent in two clicks, or add a ready-made one from PGB Marketplace and review it before it runs.
Write good instructions for an agent
How to word the Instructions box so an agent does the right job, stays inside its limits and tells you when it is unsure.
Choose tools and understand which need approval
The list of tools an agent can use, what each does, and which ones always pause for your approval.
Start agents manually, on a schedule or from a webhook
Run an agent by hand, set it to repeat with the schedule builder, or give it a private web address that starts a run.
Set budgets and know when a run will stop
Put a dollar limit on each run and each month, and learn the other limits that stop an agent from running away.