Pagination

Credicorp uses cursor pagination, not page numbers. Cursors are immune to the "shifting window" problem of offsets: inserting a record while you page through doesn't skip or duplicate rows. Every list across the API — from applications to events — behaves identically.

The list envelope

A list response is an object with object: "list", the page in data (newest first), and a has_more flag. The url echoes the resource path.

{
  "object": "list",
  "url": "/partner/v1/applications",
  "has_more": true,
  "data": [
    { "id": "app_8Kd2c9Qm", "object": "application", "status": "funded" },
    { "id": "app_7Jc1bY6n", "object": "application", "status": "approved" },
    { "id": "app_6Hb0aX5m", "object": "application", "status": "in_review" }
  ]
}

Query parameters

These parameters are accepted by every list endpoint. starting_after and ending_before are mutually exclusive.

Endpoint-specific filters — such as status, created_after or reference on applications — compose freely with the pagination parameters and persist across pages.

Paging forward

To walk the whole collection, take the id of the last object in data and pass it as starting_after on the next call. Repeat while has_more is true.

# Page 1 — newest 20
curl -s "https://hub.credicorp.co.uk/partner/v1/payments?limit=20" \
  -H "Authorization: Bearer $TOKEN"

# Page 2 — start after the last id from page 1
curl -s "https://hub.credicorp.co.uk/partner/v1/payments?limit=20&starting_after=pay_5Gd9cQ2v" \
  -H "Authorization: Bearer $TOKEN"

Auto-pagination in the SDKs

Every SDK wraps the cursor loop. Use the iterator helpers to stream an entire collection without managing cursors by hand — they fetch the next page lazily as you consume the current one.

// autoPagingIterator transparently fetches each page
$payments = $cc->payments->all(['limit' => 100, 'status' => 'settled']);

foreach ($payments->autoPagingIterator() as $payment) {
    reconcile($payment->id, $payment->amount_pence);
}

// async iterator pages automatically
for await (const payment of cc.payments.list({ limit: 100, status: 'settled' })) {
  await reconcile(payment.id, payment.amount_pence);
}

# auto_paging_iter yields every object across pages
for payment in cc.payments.list(limit=100, status="settled").auto_paging_iter():
    reconcile(payment.id, payment.amount_pence)

Paging by hand

If you manage the cursor yourself, the loop is short. Hold the last ID, stop on has_more === false:

let cursor = null, all = [];
do {
  const qs = new URLSearchParams({ limit: '100' });
  if (cursor) qs.set('starting_after', cursor);

  const res  = await fetch(`https://hub.credicorp.co.uk/partner/v1/accounts?${qs}`,
    { headers: { Authorization: `Bearer ${TOKEN}` } });
  const page = await res.json();

  all.push(...page.data);
  cursor = page.has_more ? page.data[page.data.length - 1].id : null;
} while (cursor);

Iterating safely

• Drive the loop with has_more, not data.length. A short page does not mean the end — only has_more: false does.
• Cursors are object IDs, not opaque tokens. Persist the last ID to resume a long export later from exactly where you stopped.
• Use the largest sensible limit (up to 100) to cut round-trips and stay clear of rate limits; pace bulk pulls and honour Retry-After on 429.
• Records added mid-iteration are safe. Because cursors anchor to a specific ID, a new application created while you page forward simply appears at the head of the list and won't disrupt your walk toward older records.
• For ongoing sync, prefer webhooks. Page once for the historical backfill, then keep state current from events rather than re-scanning the list.