Summary
Get an aggregate N-Shot summary for a job run: shot count, green/yellow/red agreement breakdown, override count, and the overall extraction agreement rate.
N-Shot measures extraction quality by running multiple extraction passes ("shots") and comparing their values field by field: when independent shots agree, the value is trustworthy; when they diverge, the field needs attention. The N-Shot endpoints expose these per-field comparisons for a job run, plus overrides and judge decisions (human or AI) to record which value is correct. All routes are nested under /v1/jobs/runs/{runId}/nshot/....
The summary endpoint returns aggregate statistics for all N-Shot comparisons in a run: total comparisons, agreement breakdown (green/yellow/red), override count, and overall agreement rate. Use this to quickly assess whether a schema change improved or degraded extraction quality before drilling into individual comparisons.
- Green — all shots produced the same value (high confidence)
- Yellow — partial agreement between shots (majority value exists but not unanimous)
- Red — no agreement between shots (each shot produced a different value)
/v1/jobs/runs/{runId}/nshot/summaryRequest
Response
Response fields
Response
{
"run_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
"shot_count": 3,
"total_comparisons": 420,
"green": 374,
"yellow": 32,
"red": 14,
"overridden": 6,
"agreement_rate": 0.8905
}Errors
Error responses