Criterion Runs - Verial

A Criterion Run is the result of running a single Criterion against a Task Run. The verification engine creates one Criterion Run per criterion when a task run completes.

Endpoints

Method	Endpoint	Description
`GET`	`/criterion-runs?task_run_id={task_run_id}`	List criterion runs for a task run
`GET`	`/criterion-runs/{id}`	Get criterion run details

Criterion Runs are read-only. They are created automatically when you complete a task run (see the Task Runs page and the v1 completion flow).

Criterion Run Object

Field	Type	Description
`id`	string	Unique identifier
`task_run_id`	string	Parent Task Run
`criterion_id`	string	Source Criterion
`passed`	boolean	`true` if the criterion passed
`score`	number	Score (`0.0` to `1.0`)
`details`	string \| null	Human-readable explanation of the result
`evidence`	object \| null	Structured evidence payload (field-level results, matched messages, transcript snippets). Shape depends on the criterion’s check type
`started_at`	datetime	When verification started
`completed_at`	datetime \| null	When verification finished

HTTP Example

curl "https://api.verial.ai/criterion-runs?task_run_id=tr_abc123" \
  -H "Authorization: Bearer $VERIAL_API_KEY"

{
  "data": [
    {
      "id": "crun_01H...",
      "task_run_id": "tr_abc123",
      "criterion_id": "crit_01H...",
      "passed": true,
      "score": 1.0,
      "details": "Appointment resource matched; all 2 field assertions passed.",
      "evidence": {
        "field_results": [
          {
            "path": "participant.0.actor.display",
            "expected": "Dr. Rivera",
            "actual": "Dr. Rivera",
            "passed": true
          }
        ]
      },
      "started_at": "2026-04-20T17:10:02.000Z",
      "completed_at": "2026-04-20T17:10:04.000Z"
    }
  ],
  "next_cursor": null
}

SDK Example

// List criterion runs for a task run
const runs = await verial.criterionRuns.list({ taskRunId: "tr_abc123" });

for (const run of runs.data) {
  console.log(`${run.criterionId}: ${run.passed ? "PASS" : "FAIL"} (${run.score})`);
  if (!run.passed) console.log(run.details);
}

// Fetch one criterion run with full evidence
const detail = await verial.criterionRuns.get({ id: "crun_01H..." });