Skip to main content
POST
Create a workflow-eval run against the current workflow draft. A run can execute one saved eval, every eval for one block, or every eval in the workflow. The canonical route is flat: send workflow_id in the request body and optionally narrow execution with scope. If scope is omitted, every saved eval in the workflow runs. The response is a run resource. Use its id with the run-id-first endpoints: Get Workflow Eval Run, List Eval Run Results. The request body has a workflow context and an optional scope:
  • omitted scope - run every saved eval in the workflow.
  • scope.type = "single" - run one saved eval by eval_id.
  • scope.type = "block" - run every saved eval for one block by block_id.

Authorizations

Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

Create a workflow eval run. Provide a workflow_id, and optionally narrow execution with scope to a single eval or one block. Omit scope to run every saved workflow eval.

workflow_id
string
required
scope
WorkflowEvalRunSingleScope · object

Run one saved workflow eval in the workflow.

Response

Successful Response

A batch execution of a workflow's evals, with overall lifecycle, timing, and pass/fail counts.

id
string
required
workflow_id
string
required
workflow_version_id
string
required
trigger
EvalRunTrigger · object
required
lifecycle
PendingWorkflowEvalRun · object
required

The eval run has been created but execution has not started.

timing
WorkflowEvalRunTiming · object
required
total_evals
integer
required
target
EvalRunBlockTarget · object | null

Public workflow-eval target.

The storage layer remains block-scoped today, but the API shape names the tested entity explicitly so workflow-level targets can be added later.

eval_id
string | null
counts
BlockEvalBatchExecutionCounts · object

Aggregate counts for a batch of block-eval runs.

Each individual run contributes to exactly one lifecycle_counts bucket, and additionally to one outcome bucket when lifecycle_counts.completed is incremented.

freshness
EvalRunFreshness · object

Compatibility envelope only. WorkflowEval.freshness is the authoritative read-time staleness verdict for saved eval definitions.