Skip to content

Trace message

Note: The trace message component is currently experimental and may undergo changes in future releases.

The trace message represents a meaningful step or phase in an AI activity trace. It communicates whether an activity is running, completed, or failed and can reveal additional details on demand.

Usage

Use trace messages to make multi-step AI activity transparent without exposing every internal event. Show one message for each user-relevant step or phase and arrange related messages in execution order.

  • Use a concise label that describes the activity.
  • Set state to running, completed, or failed to communicate its current status.
  • Set icon to customize the completed activity marker. It defaults to element-record-filled.
  • Project optional details such as Markdown, metadata, or tool output into the component. Users can expand and collapse this content from the trace header.
  • Use expanded to control the initial disclosure state or to synchronize it with application state.

Avoid using trace messages for low-level implementation details that do not help users understand the AI's progress or result.


Except where otherwise noted, content on this site is licensed under MIT License.