| Filter query syntax | The filter search bar and field:value query syntax, with thread, root run, single run, and any run scopes. | Filtering with legacy function-style filters. |
| Custom dashboard chart editor | The current chart editor, with templates, ratios, per-series filters, grouping, breakdowns, and chart drilldowns. | Creating custom dashboards and charts with custom dashboards (legacy), and prebuilt dashboards. |
| Trajectory view | The deduplicated conversation across turns, model calls, tools, and subagents. Also powers the Markdown export of a thread and parsed messages in experiment run details. | Trace inspection and message rendering for a single run. |
| Trajectory evaluators | Online evaluators that use the trajectory variable to score the full agent path, including child model and tool calls. | Run evaluators and thread evaluators that read root-run messages. |
| Threads in annotation queues | Adding a thread to an annotation queue from the thread view, and reviewing a thread item, which displays its trajectory. | Run-level annotation queues. |
| Threads in datasets | Adding selected threads, including thread items from an annotation queue, to a dataset as conversation examples. | Adding runs to a dataset. |
| Thread automation rules | Creating thread rules that add threads to annotation queues or datasets, or send webhooks. Backfilling a thread rule over historical data. | Run automation rules. |
| Public thread sharing | Sharing a whole thread with a public link. | Sharing a single trace with a public link. |
| Large trace actions | The paginated trace tree for large traces, downloading a whole trace as JSON, and deleting a whole trace from the trace details page. | Viewing traces, downloading individual runs, and existing deletion workflows. |
| LLM Gateway usage | The Usage dashboard: spend summary, charts, breakdowns, filters, and drilldowns. | Gateway routing and policies. |
| Agent trace volume | Trace-count sparklines on agent overview cards and the trace volume chart on an agent’s overview. | Agent deployment, management, and other agent statistics. |
| Feedback in bulk exports | Exporting the feedbacks field, which holds feedback keys and comments. | Bulk exports of every other field. |
