Receipts and verification

A live event view is presentation. An engine-issued receipt identifies the run and its available evidence; its presence alone does not prove a sealed journal exists. Refusal before execution, queued cancellation and interrupted ownership can leave no complete runtime trace.

const run = await nika.run('workflow.nika')
const result = await run.result()

if (!result.receipt) throw new Error('run did not issue a receipt')

await auditStore.write({ runId: run.id, receipt: result.receipt })

const verdict = await nika.traceVerify(result.receipt)
if (!verdict.verified) {
  throw new Error(verdict.output ?? verdict.reason ?? 'verification failed')
}

NikaReceipt is an opaque readonly record. The SDK never constructs it, extracts a path from it, enriches it by reading a workflow, or implements cryptography. It asks the authoritative engine to verify the object unchanged. The runtime writes and seals trace evidence. Verification reads what already exists; it never creates a journal or seals an incomplete one. Paused results can carry evidence for the completed observation leg while the durable job remains resumable. Local verification can return verified or invalid. The remote endpoint currently returns a typed unavailable verdict with reason trace_journal_unavailable because it has no path-free journal authority. The diagnostic route exists; this is not a 404. CLI trace inspection remains available for operators:

nika trace show .nika/traces/<run>.ndjson
nika trace replay .nika/traces/<run>.ndjson
nika trace export .nika/traces/<run>.ndjson

Replay reprojects recorded events; it never reruns the workflow.

Test and trace

Keep goldens and proof verification distinct.

Run receipts pattern

Carry receipts through application boundaries.

Overview

Start

Project

Local API

Runtime

Remote API

Operations

Reference

Receipts and verification

Test and trace

Run receipts pattern