Identify the workflow’s output unit
A seller chooses a product-listing preset and looks for the familiar control for one, two or four images. Its absence does not necessarily mean the interface failed. Before leaving the preset or repeating the run, establish whether its documented output unit is one selectable candidate or a set of complementary assets.
Kittl documents both selectable-count workflows and multi-output presets such as Product listing. The latter produce multiple outputs per run without a count picker. The Generate control shows the total token cost, and individual result tiles can finish at different times. Treat those as parts of the same requested set until the actual workflow establishes otherwise.
Write a deliverable inventory before generating
| Brief requirement | Evidence to inspect | Decision |
|---|---|---|
| One hero candidate | Preset description and actual outputs | Choose a route that matches the job |
| Listing asset set | Promised roles and output tiles | Track every required deliverable |
| One tile finished first | Remaining actual tile states | Do not declare the whole set complete |
| Another run proposed | Missing assets, failures and total quote | Do not repeat blindly to obtain a hidden count control |
Apply the distinction to a fictional lamp launch
A merchant needs a front view, a detail treatment and a lifestyle candidate for a lamp. Compare the real preset's described deliverables with that brief before any generation. If the preset creates a different set, change the production plan rather than assuming every output is interchangeable. A generated promotional panel does not replace a real specification image simply because both are listing assets.
After an authorized run, reconcile each actual output with its intended role and status. Preserve the correct lamp source and identify which candidate depicts the right switch, head shape and finish. If the lifestyle tile completes first, its availability does not establish that the detail asset is ready. This example reports no Kittl run, tile count or token charge.
Approve the set by role
Review each actual asset independently and keep unsupported outputs out of the final listing. A convenient bundle can include a strong hero and an inaccurate detail image. Record accept, revise or hold per deliverable rather than applying one approval to every tile.
The multi-page record guide addresses data mapping in another tool. The channel image-role guide explains destination-specific use. This decision concerns the fixed multi-output production unit of the chosen Kittl preset, before those handoffs.
Close the run only when every required role has a known output and review status. CartVisual does not configure Kittl presets or infer their total charge from its own credits.