payouts.batches

A payout batch is one disbursement run, frozen. Building it moves eligible ledger rows onto the batch so nothing else can claim them; exporting renders the file; confirming records what the rail actually did.

const batch = await boomin.payouts.batches.create({
  periodStart: "2026-08-01",
  periodEnd: "2026-09-01",
});

const accepted = await boomin.payouts.batches.export(batch.id);
await boomin.operations.wait(accepted.operation, { timeout: 120000 });

const exported = await boomin.payouts.batches.retrieve(batch.id);
console.log(exported.downloadUrl);

Method	Route	Scope
create(params, options)	POST /payouts/batches	payouts:write
list(params, options)	GET /payouts/batches	payouts:read
retrieve(id, options)	GET /payouts/batches/{id}	payouts:read
export(id, params, options)	POST /payouts/batches/{id}/export	payouts:write
confirm(id, params, options)	POST /payouts/batches/{id}/confirm	payouts:write
cancel(id, params, options)	POST /payouts/batches/{id}/cancel	payouts:write

Ids are prefixed pob_.

The payout batch object

{
  "id": "pob_...",
  "object": "payout_batch",
  "rail": "csv_batch",
  "status": "exported",
  "currency": "usd",
  "periodStart": "2026-08-01",
  "periodEnd": "2026-09-01",
  "itemCount": 12,
  "totalAmountCents": 48500,
  "exportFileKey": "payouts/…/pob_….csv",
  "exportFormat": "paypal_payouts_csv",
  "externalBatchRef": null,
  "exportedAt": "2026-08-03T06:20:11.104Z",
  "confirmedAt": null,
  "completedAt": null,
  "error": null,
  "createdAt": "2026-08-03T06:19:44.201Z",
  "items": [ … ],
  "downloadUrl": "https://…"
}

(Raw HTTP responses spell these period_start, total_amount_cents, and so on; the SDK camelCases every response key.)

items accompanies create and retrieve. downloadUrl appears on retrieve only — see where downloadUrl lives.

Note the batch total is totalAmountCents (wire total_amount_cents), not *Minor: the batch is a single-currency object and this is a pre-existing physical field name. Rule economics use perUnitMinor / bonusMinor.

Statuses

draft exporting exported submitted reconciling completed partially_paid failed canceled

Batch items

{
  "id": "2d7c12ab-…",
  "payoutId": "…",
  "entityId": "…",
  "userId": null,
  "recipientHandle": "creator@example.com",
  "amountCents": 2500,
  "currency": "usd",
  "status": "pending",
  "externalItemRef": null,
  "failureReason": null,
  "paidAt": null
}

Item id is a bare uuid, not a prefixed id — it is what confirm names in results[].item. Item statuses are pending processing paid failed returned canceled.

create

Synchronous. Freezes the eligible rows onto a new batch and resolves the batch plus its items and skipped. Answers 201.

const batch = await boomin.payouts.batches.create({
  rail: "csv_batch",           // optional — omit to use the brand's default rail
  periodStart: "2026-08-01",   // optional
  periodEnd: "2026-09-01",     // optional
});

Omit both period fields to sweep every eligible row regardless of period.

It is synchronous because the build is a single database transaction whatever the item count — there is no unbounded work to make durable, and a 202 would hand back an operation id for something already finished.

Eligibility

A ledger row joins a batch when its status is pending or awaiting_account, and it is not bridged to another brand’s wallet. Bridged rows settle on the wallet rail instead — never both, never twice.

A stripe_connect batch takes pending rows only.

awaiting_account is the normal status

A row lands pending only when its recipient already has an active, payouts-enabled payout account. Since entity Connect onboarding is not yet available, expect awaiting_account — and note csv_batch batches those rows anyway. Paying people who have no Boomin-side payout account is what the CSV rail is for.

skipped is a count

const batch = await boomin.payouts.batches.create({ periodStart, periodEnd });
console.log(batch.itemCount, batch.skipped); // 12  3

skipped is the number of otherwise-eligible rows dropped for want of a recipient handle — no email on csv_batch, no onboarded Stripe account on stripe_connect. Those rows stay eligible and will join a later batch once the handle exists.

If every eligible row is skipped, the build answers payout_batch_empty (409) carrying the same count.

No rail configured

import { PayoutRailRequiredError } from "@boomin/sdk";

try {
  await boomin.payouts.batches.create({ periodStart, periodEnd });
} catch (err) {
  if (err instanceof PayoutRailRequiredError) {
    // configure one — nothing is auto-provisioned
  } else throw err;
}

Every “not configured” path — no rail of the named kind, no default, a default nothing can batch, a disabled rail — collapses into that one code, because from your side they are one problem with one fix. See why nothing is auto-created.

export

202 + an operation. Writes the rendered file to storage.

const accepted = await boomin.payouts.batches.export("pob_...");
// { batch: "pob_...", status: "exporting", operation: "op_..." }

const operation = await boomin.operations.wait(accepted.operation, { timeout: 120000 });
if (operation.status !== "succeeded") throw new Error(`export ${operation.status}`);

batch and operation are id strings, never embedded objects — the same 202 contract payouts.exportCsv answers. One export contract, not two.

Repeating the call replays the same operation, and even a re-run writes the same storage key: one batch can never produce two artifacts.

Where downloadUrl lives

Not in the 202. It is minted on read:

const batch = await boomin.payouts.batches.retrieve(accepted.batch);
console.log(batch.downloadUrl);

The URL is presigned and short-lived. Returned once from the mutation it would already be expiring by the time an operator opened it, and could not be re-obtained without re-exporting. On retrieve it is regenerated every time.

downloadUrl is null on a batch that has an exportFileKey when presigning credentials are unavailable — the file exists but was not delivered. Treat that as a failure rather than writing a zero-byte file.

confirm

202 + an operation. Records the outcome of the disbursement you performed.

const accepted = await boomin.payouts.batches.confirm("pob_...", {
  externalBatchRef: "PAYPAL-2026-08",
});
await boomin.operations.wait(accepted.operation);

With no results, every item settles as paid. To report per-item outcomes:

const batch = await boomin.payouts.batches.retrieve("pob_...");

await boomin.payouts.batches.confirm(batch.id, {
  externalBatchRef: "PAYPAL-2026-08",
  results: [
    { item: batch.items[0].id, status: "paid" },
    { item: batch.items[1].id, status: "failed", reason: "recipient email bounced" },
    { item: batch.items[2].id, status: "returned", reason: "account closed" },
  ],
});

Param	Meaning
externalBatchRef	Your rail-side batch id, ≤ 200 chars. Also the retry key — see below.
results[].item	A batch item id (bare uuid) from batch.items
results[].status	paid | failed | returned
results[].reason	Optional, ≤ 500 chars

Up to 1000 results. Naming an item that is not in this batch is a typed 400 with param: "results" — it never partially applies.

Why confirm is an operation

The work is unbounded. Item count has no cap, and items settle one at a time — at least one round trip each, plus a guarded wallet debit each when the rail is walletFunded. A four-figure batch would exceed a Worker’s subrequest budget and die mid-settlement with some items paid, some not, and no handle to resume. Under the operation kernel every leg is fenced and keyed on the item id, so a retry finishes rather than restarts.

Repeating a confirm with the same externalBatchRef replays one operation, so an operator’s retry after a timeout cannot settle the run twice.

cancel

Synchronous. Unfreezes a batch that has not settled and returns its rows to the eligible pool.

const canceled = await boomin.payouts.batches.cancel("pob_...");
console.log(canceled.status); // "canceled"

list and retrieve

const page = await boomin.payouts.batches.list({ limit: 20 });
const batch = await boomin.payouts.batches.retrieve("pob_...");
console.log(batch.items, batch.downloadUrl);

list accepts limit (1–100, default 20) and startingAfter. Batches are few and already ordered, so paging is applied in memory — but it is applied.

retrieve returns the bare batch plus items and downloadUrl alongside.

Errors

Code	HTTP	When
payout_rail_required	409	No active rail of the requested kind, or no usable default. → PayoutRailRequiredError
payout_batch_empty	409	No settle-able row for this rail and period. → PayoutBatchEmptyError
payout_batch_conflict	409	A concurrent build raced this one. → PayoutBatchStateError
payout_batch_not_exportable	409	Wrong status for export. → PayoutBatchStateError
payout_batch_not_confirmable	409	Wrong status for confirm. → PayoutBatchStateError
payout_batch_not_cancelable	409	Wrong status for cancel. → PayoutBatchStateError
invalid_request	400	results names an item not in this batch.
payout_batch_not_found	404	Unknown, malformed, or another tenant’s batch id.

On any PayoutBatchStateError, read the batch and look at status — the state machine refused the verb, and the current status says why.

From the CLI

npx @boomin/cli payout batches create --period-start 2026-08-01 --period-end 2026-09-01
npx @boomin/cli payout batches list
npx @boomin/cli payout batches show pob_...
npx @boomin/cli payout batches export pob_... --out payouts.csv
npx @boomin/cli payout batches confirm pob_... --external-batch-ref PAYPAL-2026-08
npx @boomin/cli payout batches cancel pob_...

export and confirm poll their operation to a terminal status by default; --no-wait returns the 202. Full flag table: CLI reference.