Create Ecommerce Product Images from One Photo: 5 Steps

Use the uploaded product photo as the source of truth. Create a clean product
reference image for use in later ecommerce image generation.

Preserve the actual product: category, silhouette, proportions, structure,
color, materials, logo, label, packaging and all visible identifying details.
Do not redesign it, replace it with a similar item or invent missing features.

Remove unrelated background objects, clutter and distracting reflections.
Place the complete product on a pure white or light neutral background with
soft, even lighting and a subtle, natural contact shadow. Keep the edges clear
and the materials realistic. Do not crop the product.

Do not add promotional text, badges, decorations, accessories, buttons, ports,
openings or packaging details that are not present in the source.
Preserve existing product text and logos rather than rewriting them.

If several reference photos are supplied, they show the same SKU. Use them
together to preserve its appearance. Do not invent a back view, interior or
hidden detail that the photos do not establish.

Output one clean product reference image, not a collage or presentation board.

Help me create a product brief for an ecommerce image set. I am not a designer.
Use the uploaded product images and the information below.

Sales channel: [Amazon / Shopify store / another marketplace / social ads]
Target market: [country or region]
Image copy language: [for example, US English]
Number of images required: [number]
Image purpose: [main listing image / gallery images / ads / a complete set]
Aspect ratio or dimensions: [requirements, if known]
Product name:
Brand:
Product category:
Confirmed selling points:
Confirmed specifications, ingredients or materials:
Target customers:
Typical uses:
Price positioning: [budget / mid-range / premium / not specified]
People in the images: [required / not allowed / suggest whether useful]
Type of person or interaction, if needed:
Faces visible: [yes / no / not applicable]
Added text in images: [yes / no]
Things to exclude:
Other requirements:

Organize the brief into three clearly separated sections:
1. Confirmed facts from my information and clearly visible details in the photos.
2. Creative suggestions that need my approval.
3. Missing information marked "Needs confirmation."

Do not invent capacity, dimensions, ingredients, materials, country of origin,
certifications, patents, performance claims, sales rankings, awards, discounts,
prices or warranty terms. Do not infer a specification from appearance alone.

Include the product appearance that must remain unchanged, the intended buyer,
the sales channel, the image roles we may need, and the language and tone of copy.
Use the target market's spelling and the verified units I provide. Do not
convert or round specifications unless I ask you to.

Do not write image-generation prompts yet. First give me the brief to review
and a short list of questions that would materially improve it.

Revise the product brief with these corrections:
[Paste the corrections here.]

Remove unsupported claims. Keep unresolved facts in a separate
"Needs confirmation" list. Do not include them in customer-facing copy.
Keep confirmed product details unchanged and return the complete revised brief.

Act as an ecommerce art director, product photographer and image planner.
Use the approved product brief and uploaded product references to design a
coordinated image set. I will generate the images individually in Mireka.

Create the number of images specified in the brief. Give each image a distinct
purpose based on the product, audience, sales channel and buying questions.
Do not simply repeat the same product with a different background.

Before planning, verify:
1. The references show the same SKU, not different versions.
2. Product shape, proportions, color, materials and packaging stay fixed.
3. Confirmed facts, creative suggestions and unknowns remain separate.
4. The plan distinguishes the main listing image from gallery or ad images.
5. Aspect ratio and composition fit each intended placement.
6. No unsupported specification, feature or performance claim is introduced.
7. Props and accessories cannot be mistaken for items included in the purchase.
8. Usage scenes match a confirmed use of the product.
9. People appear only when permitted and useful to explain the product.
10. Text follows the chosen language, tone and approved claims.
11. Images complement each other rather than repeat the same message.
12. Every proposed image can be made from the available reference evidence.

Product consistency rules:
- Start every image prompt with a full description of the product details
  that must remain unchanged. Do not write only "same product as before."
- Use the clean product reference for every image.
- Keep existing logos and label text accurate. Do not redesign the packaging.
- Do not invent unseen internal parts, rear features, bundled items or uses.

Placement rules:
- For a main listing image, follow the requirements of the specified sales
  channel and category. If a pure white product-only image is required,
  do not add a lifestyle scene, text overlays, badges or unrelated props.
- Treat lifestyle scenes and information graphics as separate gallery or
  advertising candidates, subject to the destination's rules.

People rules:
- If people are not allowed, include no people, faces, hands or body parts.
- If hands are allowed but faces are not, specify a hands-only composition
  with no visible face, eyes, nose or mouth.
- Keep anatomy and product scale natural. Use fictional, age-appropriate
  people rather than real-person or celebrity likenesses.

Text rules:
- If added text is not allowed, include no new headings, badges, icons or
  promotional copy. This does not mean removing the real product label.
- If text is allowed, use short, approved copy in the requested language.
  Specify its exact wording and placement. Do not invent claims.
- Keep unknowns and planning placeholders out of final image prompts and
  customer-facing copy. List unresolved questions separately.

Return:
A. A concise product appearance specification that applies to the full set.
B. An image-plan table with: image number, buyer question, intended placement,
   composition, background, lighting, props, people, exact copy and review notes.
C. A complete, standalone prompt for every image, numbered to match the table.
   Include product preservation, composition, setting, light, materials,
   permitted props, people, copy placement and exclusions in each prompt.
D. Any remaining questions that must be resolved before generation.

Keep the set visually coordinated, but vary compositions where that helps
communicate different information. Output the plan and prompts as text;
do not combine all images into one collage.

Compare these four inputs:
1. The original product photo.
2. The approved clean product reference.
3. The generated result.
4. The prompt used for that result.

The problem I noticed is:
[Describe the problem.]

Return five things:
1. A short list of visible differences that matter for this product image.
2. Whether the cause should be fixed in the reference, product brief,
   image plan or this individual prompt.
3. What must stay unchanged in the next attempt.
4. A complete revised prompt I can paste into Mireka.
5. A focused checklist for reviewing the new result.

Do not fix one issue by redesigning the product or inventing new claims.
Preserve the successful parts of the current composition unless they conflict
with the correction. If a detail cannot be established from the inputs,
ask for a clearer reference rather than guessing.