Skip to main content

Triage Breakdown

A classification-backed breakdown of the review queue: counts by reason, field, document, assignee, and tag, each group carrying a per-group reason mix.

The triage endpoint returns the review queue's shape rather than its items: how the held fields break down byReason, byField, byDocument, byAssignee, and byTag. Each group carries a per-group reason mix, so you can see not just how many items a field or document has, but why they are held — disagreement, low confidence, a failed check, a missing value. It is what a review dashboard reads to draw section headers and prioritize work.

Triage honors the same scope filters as the queue list (pipeline_id, schema_id, document_id, stage_id, trigger, older_than_hours, include_held, assignee, and the free-text q), so you can scope the breakdown to one pipeline or one schema. It intentionally does NOT honor reason or tag filters — those facets stay the full switchable vocabulary so every reason and tag is represented.

Use triage to power a queue header or routing logic: the reason mix tells you which fields are mostly safe to bulk-approve versus which need a careful human look.
GET/v1/field-reviews/triage

Response

Response fields

byReasonobjectCounts per reason bucket (missing_value, disagreement, failed_check, low_confidence, likely_fine, other).
byFieldobject[]Per-field groups with count and reason mix.
byDocumentobject[]Per-document groups with count and reason mix.
byAssigneeobject[]Per-assignee groups (null owners folded into one unassigned bucket).
byTagobject[]Per-tag groups (an item joins every tag on its schema field).