Open Batch Creation
From the Creative hub, open the Batch Creation template. The template is the canonical entry-point for catalog operations — multiple sub-flows (Ghost Mannequin, White Studio) can be queued in one batch.
How-to · Batch ops
Batch Creation is Apiway's unattended-batch operating mode for catalog work — Shopify weekly drops, Amazon FBA Q4 ramp-ups, and multi-vendor marketplace onboarding. Up to 50 garments per session, sub-batches by template, runs while you do something else.
Last reviewed: .
From the Creative hub, open the Batch Creation template. The template is the canonical entry-point for catalog operations — multiple sub-flows (Ghost Mannequin, White Studio) can be queued in one batch.
Drop your garment photos into the upload area. The cap is 50 per session. The pipeline tolerates a mix of input types within one batch — hanger photos, flat-lays, on-mannequin, on-model — and segments each garment independently.
Choose Ghost Mannequin for catalog main images, White Studio for on-model PDP shots, or queue both in sequence. For a multi-marketplace operation (US + EU + JP), run White Studio twice with different model demographics.
Pick aspect ratio, resolution, pose set, and (for White Studio) AI model demographic. The settings apply to every garment in the batch — the operating mode is unattended, not click-by-click.
Click Generate. The batch runs in the background — you can close the browser, the work continues server-side. Heavy templates (Ghost Mannequin, AI Photoshoots) run on a dedicated worker pool with 10 parallel containers.
When the batch finishes, every output lands in your gallery, grouped by source garment. Review the full set, re-run any garments where the output needs a different pose or model, and download the approved set as JPG up to 4K.
A 50-garment Ghost Mannequin batch typically completes in 5–10 minutes; a 50-garment White Studio batch with a 4-pose set per SKU runs 15–30 minutes. Heavy templates run on a dedicated worker pool (10 parallel containers, 64GB RAM) so wall-clock time scales sub-linearly with batch size.
Failed items don't deduct credits — `try_refund_credits` runs server-side on failure. The successful items in the batch are unaffected. Re-submit the failed items individually after fixing the source (e.g. better lighting on the source photo) or with a different template.