Beginner4 min read

Understanding run outputs

Navigate the timeline, confidence scores, and Output Hub on any completed run.

Every completed run produces a structured set of outputs from each agent. This guide explains how to navigate the live timeline, interpret confidence scores, and get the most from the Output Hub.

The live timeline

While a run is in progress, the Timeline tab shows events in real time. Each agent appears when it starts and updates when it finishes. You can see the agent's status (running, completed, failed) and click any completed agent event to read its output immediately — you don't need to wait for the whole run to finish.

Confidence scores

Each agent output includes a confidence score from 0.0 to 1.0. The score reflects how certain the agent is that its output is accurate and complete given the available data.

0.90 – 1.0High confidence. The agent had sufficient data and produced a reliable output.
0.85 – 0.89Moderate confidence. Review the output and verify key assumptions before acting on it.
0.0 – 0.84Low confidence. The agent likely lacked key data. Improve your brief or add a data tool to this agent.

The Output Hub

Click the Output Hub tab on any completed run. This shows all deliverables from the run in one place. Each output card shows the agent role, output type (document, code, decision, recommendation), confidence score, and a preview of the content. Click any card to read the full output.

Use the download buttons to export individual outputs as Markdown or plain text, or export the entire run as a ZIP containing all outputs. The run summary card at the top shows the overall outcome, total agent count, and run duration.

Tip:Start reading from the last agent — it usually produces the synthesised, executive-level output that references the work of all previous agents. Then drill into individual agent outputs for detail.
Understanding run outputs — Forge Guides — Forge