ShopixHelp Center

Making detail images

Pick the detail modules you want; the AI writes a plan for each image, you confirm, and the whole set is generated. New accounts get one full page free.

9 steps · 4 minCost: 50 credits per image by default; plans cost 5 credits per module after your first purchase; a new account’s first full page is free.
Open Detail images

What you bring

The result

Two modules — hero visual and core selling point — from one product photo.

Before you start

  • A few photos of the same product (front, side, details)
  • Selling points and specs: capacity, size, material, power…
  • Your marketplace and copy language

Steps

  1. Step 1 of 9

    Basic or Pro

    “Basic”: write a few lines and each image is planned for you. “Pro” runs your instructions as written, image by image if you like (“image 1 / image 2 …”). Start with Basic.

  2. Step 2 of 9

    Upload product images

    Add them under “Product images”; different angles of the same product work best. More isn’t better — upload only what you need.

  3. Step 3 of 9

    Write the requirements

    Follow the prompt “My product is …, key selling point is …”, and add where you sell and the copy language. Stuck? “Draft it for me” writes a first version.

    TipAny number that must appear — capacity, size, wattage — has to be in the requirements. The AI cannot guess it.

  4. Step 4 of 9

    Pick a detail version

    “Detail version” sits at the top of “Settings”, from most natural to most polished: “Real-shot” looks like a real photo, true to the product and unfussy, and renders on GPT Image 2.5 or Nano Banana 2.1 only, so the model switches when you pick it; “Practical” is accurate with clear information; “Cinematic” is more polished. Every marketplace offers all three and one is required; with a marketplace picked, its rules are applied on top of the version. Turn on “Set mode” next to them to keep all images consistent.

  5. Step 5 of 9

    Pick the marketplace, style and language

    Under “Marketplace” choose international, domestic or generic and the AI follows that marketplace’s image rules; “Style direction” on “Auto” follows your product and marketplace. The language of the text on the images is “Output language” under “Settings”.

  6. Step 6 of 9

    Pick detail page modules

    Each module makes one image by default: hero visual, core selling point, usage scene, product detail, size, after-sales guarantee… Pick them in the order your page should read.

    Save creditsTry 2–3 key modules first, then add the rest.

  7. Step 7 of 9

    Create the plan

    With the modules picked, press “Create plan” and the AI writes the visual and copy for each module; the line under the button shows the estimate. Until a version is picked, “Pick a detail version to continue” above the button jumps back up to it. On an account that has never purchased, the button reads “Generate detail long page” instead: the first full page is free, with no plan step.

  8. Step 8 of 9

    Check the plan, then generate

    The “Detail page plan” on the right lists each image; open one to edit its title and description. When it looks right, press “Generate N images”, which shows the credits.

    NoteStay on the page while the plan is open and generate right after editing. Your requirements are saved as a draft.

  9. Step 9 of 9

    Download, or keep refining

    Once images are rendering you can leave; finished images are in History. Each has “Edit” and “Download”; “More” next to “N in this batch” can “Stitch into one image” for a long detail page.

Not happy with the result?

I can’t find “Create plan”
A new account’s free first page and Pro mode generate directly without a plan — press “Generate detail long page” or Pro’s generate button.
Numbers or specs are wrong
Put exact specs in the requirements; fix a single number with “Edit”.
Images don’t look like one set
Turn on “Set mode” and generate again.
Images look too retouched, not like real photos
Set “Detail version” to “Real-shot”: natural, true to the product, rendered on GPT Image 2.5 or Nano Banana 2.1.
The model switched to GPT Image 2.5 by itself
“Real-shot” renders on GPT Image 2.5 or Nano Banana 2.1 only; pick “Practical” or “Cinematic” to use another model.
The copy is in the wrong language
Set “Output language”.

Related questions

Did this lesson help?