Examples & practical guide
Multi Image Kontext Max: combine two references
Multi Image Kontext Max is an experimental two-image composition app. Upload two references and explain how their contents should work together. A useful result needs more than attractive rendering: important subject details, believable integration, and the requested relationship between the sources all need to survive.
Two references · one composition
Assign a job to each image before you write the prompt
Image 1: the thing that must remain recognizable
Choose the person, object, or base scene that anchors the result. A simple source makes it easier to identify what should survive the edit. Avoid a collage containing several candidates for the same role.
Image 2: the context you want to borrow
Supply the environment, material treatment, or second subject. Say which details to use and which to ignore. Multi Image Kontext Max combines two inputs; it is not a layer editor with locked pixels or a promise of exact copying.
Three worked composition briefs
These are suggested input pairs and prompts, not reported test outputs.
Product in a new environment
- References
- Image 1: your ceramic mug on a plain surface. Image 2: an uncluttered kitchen counter photographed at a similar height.
- Prompt
- “Place the mug from image 1 on the counter in image 2. Keep the mug’s handle, shape, and glaze color. Match the kitchen light and add a natural contact shadow. Do not add other mugs.”
- Inspect
- Compare handle position, rim shape, glaze markings, and scale against the source. A persuasive mockup can still misrepresent the product, so do not publish altered specifications as a listing photo.
Person within an owned location photo
- References
- Image 1: a consenting adult’s portrait. Image 2: an empty workspace you have permission to use.
- Prompt
- “Create a waist-up portrait of the person in image 1 standing in the workspace from image 2. Preserve their facial features and hairstyle. Use the room’s lighting and keep furniture behind them.”
- Inspect
- Check face drift, shoulder edges, furniture overlap, and light direction. Present the output as a composite when it might otherwise imply that the person was photographed in that room.
Original illustration with a new palette
- References
- Image 1: your own simple character drawing. Image 2: a color-and-texture reference that you own.
- Prompt
- “Redraw the character from image 1 using the muted palette and paper texture of image 2. Keep the character’s silhouette and expression. Do not include objects from image 2.”
- Inspect
- Separate palette transfer from unwanted content transfer. If objects leak into the result, simplify the reference rather than adding a long list of contradictory instructions.
Use a three-part acceptance check
- Identity: compare the details that make the subject itself, not merely a similar subject.
- Integration: inspect contact shadows, perspective, lighting, and the boundary where the references meet.
- Brief: confirm that requested additions happened and protected elements did not change.
Upload both references, state their roles, choose the intended aspect ratio, and check the estimate before generating. If the composition fails, change one instruction or replace one incompatible reference. Strongly different camera angles are often better addressed by choosing a better input pair.
There is no quality percentage or guaranteed advantage over Multi Image Kontext Pro stated here. Compare the same brief if you want to assess both. For a genuine list of references rather than two slots, use Multi Image List.
Before you create
Multi Image Kontext Max: common questions
How many images does Multi Image Kontext Max accept?
This app requires two images in separate input slots and a text prompt. The word Multi does not imply an unrestricted image list. Use Multi Image List when you need the list-input workflow.
Is Max always better than Pro for combining images?
Not necessarily for your particular sources and instructions. This guide makes no measured quality or speed comparison. Compare the same input pair and brief, then assess the details that matter to your project and the estimate shown before each run.
Can I rely on exact product labels and logos?
No. The composition can redraw or distort small text, logos, proportions, and details. Inspect them against the source and use conventional editing when exact brand artwork or product information must remain unchanged.
Choose your next step
- Multi Image Kontext Pro
Compare another two-reference composition workflow.
- Multi Image List
Organize a list of references with distinct roles.
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