Skip to main content
GET
Saved runs, newest first. The filter that earns its keep is needs_review=true: it returns only runs where the judge’s lowest confidence fell below 0.85, the upstream’s own escalation line. That is the queue a person should actually work through, instead of re-reading everything the model was already sure about.

Request

Headers

string
required
Bearer token — your standard 60db API key (sk_…), or a user JWT. There is no separate Judge credential.

Query parameters

integer
default:"25"
1–100.
integer
default:"0"
string
evaluate or extract.
string
Only runs of one saved rubric — the whole series for that rubric.
boolean
default:"false"
Only runs whose min_confidence is below 0.85.
number
Your own threshold (0–1). Overrides needs_review.

Response

array
Each row carries id, kind, rubric_id, label, summary, state_preview, min_confidence, input_tokens, escalated_questions, latency_ms, credits_charged and created_at.Rows deliberately omit the full request and result payloads — each can be up to 32 KiB. Use Get run for those.
object
total, limit, offset, has_more.

Visibility

A run stores the content it judged, so history is not workspace-public. You see your own runs; workspace owners and admins see everyone’s.

Example