Skip to main content
POST
Legacy Dataset, Evaluation, and Report endpoints are deprecated for new workflows. Use the Tables API for new dataset import, evaluation, scoring, recalculation, and reporting workflows.
Add an observability trace (or a specific span subtree) as a row in the draft dataset version for a dataset group. You can export an entire trace (anchored at the earliest root span) or choose a specific span to anchor on — in which case that span and its direct children become the dataset columns.

​
Behavior Notes

  • If no draft dataset version exists for the group, one is automatically created and seeded from the most recent published version’s rows.
  • The trace and dataset group must belong to the same workspace.
  • Use span_id to export a specific span subtree instead of the full trace root.
  • The mode field in the response indicates whether the row was created in trace mode (root anchor) or span mode (chosen anchor).

Authorizations

​
X-API-KEY
string
header
required

Body

application/json
​
dataset_group_id
integer
required

ID of the dataset group to add the trace row to.

Required range: x >= 1
​
trace_id
string
required

ID of the trace to add as a dataset row.

Minimum string length: 1
​
span_id
string | null

Optional span ID. When provided, the row is built from this span and its direct children become the dataset columns. When omitted, the full trace is added as the row.

Minimum string length: 1

Response

Trace added to draft dataset.

​
success
enum<boolean>
Available options:
true
​
draft_dataset_id
integer

ID of the draft dataset the trace row was added to.

​
mode
enum<string>

"trace" when the row was added from the full trace, or "span" when added from a specific span.

Available options:
trace,
span