Paginating results

Collections are paged with an opaque cursor. Every page carries nextCursor; pass it back to get the next one, and stop when it is null.

curl "https://fabric.inc/api/v1/brands/$BRAND/products?limit=100" \
  -H "Authorization: Bearer $FABRIC_API_KEY"

{
  "data": [{ "id": "prd_...", "sku": "SW-001" }],
  "nextCursor": "eyJpZCI6InByZF8..."
}

curl "https://fabric.inc/api/v1/brands/$BRAND/products?limit=100&cursor=eyJpZCI6InByZF8..." \
  -H "Authorization: Bearer $FABRIC_API_KEY"

nextCursor is always present — null on the last page rather than absent — so a loop can test it without distinguishing missing from empty.

There is no page number

offset, page and skip are refused with 400, not ignored.

A catalog changes while you read it. An import lands, a product is archived, and every row after the change shifts position. Offset paging then silently skips rows and returns others twice, and nothing in the response says so. A cursor points at a row rather than a position, so neither happens.

Limits

limit defaults to 50 and caps at 200. Above 200 the request is refused, not quietly clamped — asking for 1000 and receiving 200 looks identical to a catalog with 200 products in it.

Walking the whole catalog

cursor=""
while :; do
  page=$(curl -s "https://fabric.inc/api/v1/brands/$BRAND/products?limit=200&cursor=$cursor" \
    -H "Authorization: Bearer $FABRIC_API_KEY")
  echo "$page" | jq -c '.data[]'
  cursor=$(echo "$page" | jq -r '.nextCursor // empty')
  [ -z "$cursor" ] && break
done

A cursor stays valid across pages of a changing catalog. It carries a position and nothing else — no filter, no brand, no permission — so a walk cannot outlive what your credential may read.