Skip to main content

Step 1 - Configuring images in Bulk Optimisation

Generate new listing images using product images, prompts, and reference assets.

Configuring images in Bulk Optimisation

Image settings control how Sprintr creates new product images during a Bulk Optimisation run.

You can use image generation to create fresh lifestyle images, improve the visual quality of listings, add stronger product context, or create more consistent imagery across several products at once.

Image generation can be used on its own, or alongside copy and video generation.


Turning image generation on or off

In the Configure step, you can choose whether Images are included in the run.

If Images are switched on, Sprintr will generate new images for each selected listing.

If Images are switched off, Sprintr will not create new images for that run.

You can combine Images with Copy and Videos, or run image generation on its own.


Image credit cost

Images use credits.

Your run summary shows the estimated image cost before you generate anything.

The estimate is based on:

  • number of selected listings

  • number of images per listing

  • image credit cost and the model you used.

For example, if you select 3 listings and generate 1 image per listing, Sprintr will estimate 3 images in total.

Copy is included, but images use credits.

Nothing is generated until you confirm the run.


Images per listing

Use Images per listing to choose how many images Sprintr should generate for each selected listing.

For example, if you choose:

1 image per listing

Sprintr will generate one new image for each selected product.

The run summary updates automatically so you can see the estimated credit cost before continuing.


How Sprintr uses the primary product image

Each listing’s primary image is used as the main product reference for image generation.

This helps Sprintr understand what the product looks like and keeps the generated image connected to the correct listing.

For example, if the listing is for a candle set, the listing’s primary image acts as the main product reference for the new generated image.

You can change the primary product image later in the configuration where available.


Reference assets

Reference assets give Sprintr extra visual guidance.

You can add reference assets such as:

  • logos

  • labels

  • packaging

  • on-model images

  • supporting products

  • lifestyle references

  • brand assets

  • visual style references

These can be pulled from Asset Manager, uploaded during the flow, or selected from listing assets where available.

Reference assets help Sprintr understand more than just the product itself. They can guide the setting, model, branding, packaging, or supporting visual details.


Adding reference assets

In the Image generation section, use:

Add from library

to select an existing asset from Asset Manager.

Use:

Upload

to add a new asset during the run.

You can add up to 4 shared reference assets for image generation.

These shared assets apply across the selected listings, but Sprintr still adapts the generated output to each individual product.


Asset tokens in prompts

When you add a reference asset, Sprintr can insert it into the prompt as an asset token.

For example, you may see tokens such as:

  • Primary product

  • Mark model

  • another selected image or asset name

These tokens tell Sprintr which asset the prompt is referring to.

For example:

A professional lifestyle photograph featuring [Mark model] holding [Primary product] in a bright, modern setting.

This helps Sprintr understand the role of each asset in the image.


Using the suggested prompt

Sprintr can suggest a prompt based on the product and selected assets.

Use the suggested prompt when you want Sprintr to recommend a strong starting point.

You can then edit the prompt before generating.

This is useful if you know the kind of image you want, but do not want to write the full creative direction from scratch.


Writing your own image prompt

You can also write your own shared creative prompt.

The prompt should describe the kind of image you want Sprintr to create.

For example, you might describe:

  • the setting

  • the mood

  • the customer or model

  • how the product should be shown

  • the lighting

  • the background

  • whether the image should feel premium, casual, minimal, editorial, or lifestyle-led

The same prompt is applied across the selected listings, but Sprintr adapts it to each product.

For example, a shared prompt for “a bright minimalist kitchen setting” may produce different images for a mug, a candle, and a storage basket because each listing has its own product context.


Adjusting asset prominence

You can adjust how important an asset should be in the generated image.

For example, an asset may be set as:

Subtle
The asset is used as a light reference.

Standard
The asset has normal importance in the prompt.

Hero
The asset should be treated as a dominant subject.

This is useful when you want one asset to strongly shape the image, while another asset only provides supporting context.

For example, the primary product may usually be the hero subject, while a model or brand reference may be used with standard or subtle prominence.

You can also remove an asset token if you no longer want that asset to guide the image.


Image model and composition settings

The Image generation section may also show model and composition settings.

These can include details such as:

  • image model

  • orientation

  • number of images per listing

  • whether people are included

  • whether text is included

  • logo placement

For example, you may choose a 16:9 image with people included, text turned off, and no logo.

Use these settings to keep generated images aligned with the marketplace, brand style, or campaign need.


Logo placement

Where available, logo placement controls how a logo should be used in the generated image.

Options may include settings such as:

  • no logo

  • fixed placement

  • placement guided by the prompt

If you are using a logo as a reference asset, make sure the prompt clearly explains how you want it used.

For example, you may want the logo to appear on packaging, on a label, or as a subtle brand mark in the scene.


Text in image

Text in image controls whether Sprintr should include visible text in the generated image.

Use this carefully.

Some marketplaces have rules about promotional text, badges, claims, or overlays in product images.

If you are unsure, keep text off and add any required wording through the listing copy instead.


People in shot

The people setting controls whether generated images may include people. If you add an asset which is a person, this is automatically turned on.

Use People on when a person, hand, model, or lifestyle scene would help show the product in use.

Use People off when you want a cleaner product-led image without people.

This is useful for products where lifestyle context matters, such as apparel, homeware, skincare, gifting, or accessories.


Orientation

Orientation controls the shape of the generated image.

Common options may include:

  • 1:1

  • 4:3

  • 16:9

  • 9:16

Choose the orientation based on where the image will be used.

For example, square images may work well for marketplace galleries, while wider or vertical images may be better for campaigns, banners, or social-style content.


How image settings apply across multiple listings

Image settings are shared across the run, but Sprintr adapts the output for each listing.

This means you can set one shared creative direction, while each product still receives its own generated image.

For example, if you select three homeware listings and ask for a bright, premium lifestyle image, Sprintr will generate separate images for each listing using each product’s own primary image and product details.


Reviewing generated images

Generated images do not go live automatically.

After the run completes, you review each generated image in the Review step.

You can:

  • view the generated image

  • check the image title, alt text, caption, or description where shown

  • approve the image

  • skip the image

  • regenerate a replacement

  • use the image if you are happy with it

Only approved images are included when you publish or schedule the changes.


Regenerating an image

If a generated image is not right, you can regenerate it during review.

The regenerate image flow lets you create a replacement using similar controls to the image setup.

You can adjust:

  • main product reference

  • reference assets

  • prompt

  • asset tokens

  • creative direction

Sprintr then creates a replacement preview.

The original generated image is not replaced until you choose:

Use this image

You can also choose to try again.


Manual upload notes

Some connected stores may have limits on what can be automatically added or updated.

Where Sprintr cannot automatically apply an image or related media item, it will flag this clearly.

You may need to download the asset and upload it manually in the store’s media library.


Good ways to configure images

For best results:

  • choose listings that share a similar visual style or campaign goal

  • use clear reference assets from Asset Manager

  • keep prompts specific but not overloaded

  • use asset tokens so Sprintr knows which assets matter

  • make the product the hero when the listing image should focus on the item

  • use people only when they help explain or sell the product

  • keep text off unless you have a clear reason to include it

  • check marketplace image rules before publishing generated images

If different products need very different image styles, consider creating separate bulk runs.


What you should take away

Image settings tell Sprintr how to create new listing images across your selected products.

You can control the number of images, creative prompt, primary product reference, reference assets, asset prominence, orientation, people, text, and logo settings.

Images use credits, and the estimated cost is shown before generation.

Nothing goes live until you review and approve the generated images.

Did this answer your question?