Viewing Traces

The trace view supports multiple ways to explore the same trace: Tree, Timeline, and Reader (via the View dropdown), plus Metadata and Ask AI (via dedicated buttons/toggles). In any view, selecting a span opens the full span detail pane on the right.

​

Tree View

​

Open Tree View

1. Open a trace (from Traces).
2. In the trace view toolbar, open the view mode dropdown.
3. Select Tree (you can switch back the same way).

• Hierarchy-first view: Shows parent/child nesting so you can follow the trace’s call structure.
• Collapsible branches: Collapse spans with children to hide entire subtrees and reduce noise.
• Inline metrics: Duration on every row; tokens/cost badges on LLM spans.
• Fast navigation: Virtualized list stays responsive even on large traces.
• Synced minimap: Jump to spans quickly and scrub the shared session-time cursor.

Use Tree View when you want to understand structure: what called what, and where time/cost sits in the hierarchy.

​

Reader Mode

Reader Mode gives you a readable, flattened view of a trace designed for fast scanning and review.

​

Open Reader Mode

1. Open a trace (from Traces).
2. In the trace view header, open the view mode dropdown.
3. Select Reader Mode (you can switch back the same way).

• High-signal view: Shows only LLM and tool-call spans by default (other spans hidden), making the trace much easier to parse.
• At-a-glance columns: Type/model, timing, tokens/cost, and breadcrumbs per row.
• Expandable rows: Open any row to see formatted outputs; selecting still opens the full span detail pane on the right.
• Lazy-loaded outputs: Fetches only what’s on screen to keep the UI fast.
• Templates per path: Apply saved Mustache templates per span path to reshape how outputs render.
• Synced minimap: Highlights spans in view and lets you jump or scrub session time.

Use Reader Mode when you want a narrative, high-signal view of a trace focused on LLM + tool calls, without drilling through nested spans.

​

Timeline View

​

Open Timeline View

1. Open a trace (from Traces).
2. In the trace view toolbar, open the view mode dropdown.
3. Select Timeline (you can switch back the same way).

• Time-first view: Renders each span as a colored bar on a single time axis.
• Zoom controls: Use +/- to zoom in for fine-grained alignment or out for a high-level overview.
• Event markers: Span events appear as ticks inside the bar.
• Inline labels: Name + duration (and tokens/cost for LLM spans) are shown on the bar when space allows.

Use Timeline View when you’re diagnosing latency, overlaps, gaps, or long-running steps.

​

Metadata

​

Open Metadata

1. Open a trace (from Traces).
2. In the trace view toolbar, click Metadata.

• Read-only source of truth: Shows the trace metadata exactly as stored.
• JSON-first: Defaults to a JSON view with formatting and syntax highlighting.
• Search/copy friendly: Quickly find keys/values and copy raw text as needed.

Use Metadata when you need to verify context fields (inputs, run config, tags, custom attributes) without digging through spans. See Metadata.

​

Ask AI

​

Open Ask AI

1. Open a trace (from Traces).
2. In the trace view toolbar, click Ask AI.

• Automatic summary: Generates a trace summary (and deeper analysis) to orient you quickly.
• Span-aware answers: When the assistant cites spans, they show up as clickable chips that jump you to the referenced span.
• Iterate fast: Ask follow-ups about errors, unexpected outputs, tool calls, or performance.
• Resettable: Start a new chat when you want a fresh thread.

Use Ask AI when you want a quick narrative explanation or guided debugging without manually scanning the whole trace.

​

Open in SQL Editor

1. Open a trace (from Traces).
2. In the trace view header, open the Trace dropdown.
3. Select Open in SQL editor (opens in a new tab).

• Pre-filled query (trace): Starts with SELECT * FROM spans WHERE trace_id = ... ORDER BY start_time ASC.
• Pre-filled query (span): From a span’s detail header dropdown, Open in SQL editor creates SELECT * FROM spans WHERE span_id = ....

Use SQL Editor when you need precise filtering, aggregation, or auditing on the raw spans table beyond what the UI exposes. See SQL Editor.