Should a handheld bag image use a product or virtual try-on workflow?

Choose the sale item before the person

Fictional routing example, not a completed generation: a merchant sells a structured handbag with a rigid handle and detachable strap. The jacket on the model is styling. The bag is the SKU, so the first source set must establish its body, handle, strap attachments, closure and material.

A second merchant sells the jacket and wants to show its neckline, fastening and hem on a person. That is a garment-presentation task. The presence of a face is common to both images; it does not make their evidence or workflow interchangeable. Use the fashion source review for the wider garment-and-accessory context.

Select the purpose that permits the intended composition

In Product images, choose Lifestyle for a held-bag composition and add an optional model only when the interaction helps explain the bag. Supply the exact bag photographs and describe one simple hold. White-background hero and Material details are product-only purposes; choosing a person does not turn those outputs into a held-bag scene.

For a garment being sold, use Virtual try-on with the actual garment references. It creates garment imagery, not a measurement-based fitting result. Botika’s clothing and accessory FAQ is a separate vendor example: its accessory limitations do not decide CartVisual’s capabilities.

If the task requires a particular worn strap arrangement, obtain a reference showing that arrangement first. A scene brief cannot establish a hidden attachment.

Trace the handle-to-hand chain

Review the downloaded image beside the exact bag photograph. Start at one handle attachment, follow the handle through the hand and return to the other attachment. Check the opening, handle shape and hardware count before judging the pose.

Then follow the strap separately. It must not pass through the bag body, merge with a sleeve or end at invented hardware. Inspect fingers at close view, then return to the complete frame for support, apparent scale and product visibility. An unseen connection is unresolved, not verified. Use the occlusion review when hair, clothing or hands cover a defining detail.

Act on the specific failure

In the fictional handbag example, candidate A has an attractive jacket but merges two fingers into the rigid handle. Reject that grip; change to a simpler hold or use the unobstructed bag view. Candidate B keeps the visible handle but hides one strap attachment. It may serve a styling role after the other checks, but it cannot prove where that strap connects.

A request to show the interior creates a different evidence need. Obtain an actual open-state photograph rather than enlarging the generated bag or inferring its capacity from the person. These are proposed review decisions, not observed generation results.

Keep the approved role with the file

Record the SKU and edition, source filenames, selected workbench and purpose, model reference if used, exported filename and the specific interaction reviewed. Mark hidden details or unresolved contacts explicitly. Keep real dimensions and verified capacity information separate from the generated scene.

Personal model references must be authorized adult references. A generated person is not a real purchaser or a calibrated scale. Before reuse, compare the actual product and surrounding copy again; a styling image should not silently become construction or fit evidence. The presentation-versus-fit distinction explains that separate claim boundary.