Examples & practical guide
Multi Image Kontext Pro: combine two images
Multi Image Kontext Pro combines two uploaded images according to a text instruction. It is useful for exploring an object in a new setting or a subject with a borrowed visual treatment. Start with one small, testable change and inspect the result instead of assuming that “combine” means exact cut-and-paste.
A small, testable edit
Start with “keep this, borrow that”
Multi Image Kontext Pro works best as a clearly specified two-image task, not an instruction to mix everything equally. Name the subject to keep, the reference to borrow from, and the relationship between them. The app requires two image uploads and a prompt; it does not provide a brush, mask, or adjustable blending percentage.
A reusable prompt structure
Use [subject] from image 1 in [setting or treatment] from image 2. Preserve [important details]. Place it [spatial relationship]. Keep [unwanted changes] out of the result.
Replace each bracket with a concrete description. “The small red backpack” is clearer than “it” when both references contain several objects.
Pick a first experiment
A backpack on a bench
Inputs: a side-view backpack photo and a park bench photographed from a compatible angle.
Suggested prompt: “Place the red backpack from image 1 on the right side of the bench in image 2. Keep its pockets and straps. Match the outdoor light and leave the rest of the bench empty.”
Check: count straps and pockets, then inspect the contact point. A bag floating above the bench is a failed integration even when its color looks right.
A portrait as an illustration
Inputs: your portrait and an original line-and-color illustration reference.
Suggested prompt: “Illustrate the person in image 1 using the linework and limited colors of image 2. Keep the hairstyle and face shape. Use only the style, not the people or objects in image 2.”
Check: likeness and unexpected borrowed objects. A simpler illustration reference can make the instruction less ambiguous.
A lamp in a room concept
Inputs: a lamp product photo and your room photo with a clear side table.
Suggested prompt: “Place the lamp from image 1 on the side table in image 2. Preserve the lampshade shape. Keep the room layout and make the lamp a plausible size relative to the table.”
Check: proportions, furniture changes, and invented electrical details. This is a visual concept, not an installation plan or a scale drawing.
Run a useful comparison rather than a guessing loop
- First passUse one simple requested change, your intended ratio, and the clearest references available. Save the prompt alongside the output.
- Second passKeep both images fixed and clarify the single weakest relationship: placement, preserved shape, or borrowed style. Changing every detail hides what helped.
- Stop conditionAccept only when the important subject details and the composition both work. For exact labels, logos, or dimensions, finish in a conventional editor rather than trusting a similar-looking regeneration.
Know when another workflow fits better
If you need more than two references, Multi Image List exposes a list input instead of this fixed pair. To evaluate another two-image option, try the same source pair and brief in Multi Image Kontext Max; do not assume the name guarantees a better result for your particular image.
The experimental workflow can change small details, merge subjects, or ignore an instruction. Seed, aspect ratio, and output format help configure a run but do not provide exact spatial control. Check the current credit estimate for each submission and use only references you have the right to edit.
Before you create
Multi Image Kontext Pro: common questions
Do I need both images and a prompt?
Yes. This app requires Input Image 1, Input Image 2, and a prompt. Explain what each source contributes. Leaving the relationship vague makes it harder to judge whether the result actually meets your brief.
Does this app preserve untouched pixels?
No exact pixel-preservation guarantee is provided. Even elements you ask to keep may change. There is no mask or layer-lock control in this form; use a conventional editor for edits that must leave every other pixel unchanged.
How should I compare Pro and Max?
Keep the references and core instructions consistent, record the settings, and compare likeness or product details as well as overall composition. Check each app’s displayed credit estimate. Do not assume that a model name predicts the best result for every brief.
Choose your next step
- Multi Image Kontext Max
Try the same two-image brief in another app.
- Multi Image List
Use multiple references with an explicit input list.
About this guide: AI-assisted writing, checked against the tool’s available settings. Suggested setups are creative starting points, not measured performance tests. AI results vary; inspect your own output before publishing. Report a correction.
Before generating, check the credit estimate and the visibility setting shown in the tool. Use only media you have permission to edit. See plans and credits, our privacy policy and terms.
Prepare your references.
Back to the generator