To combine two images with AI, decide what each image contributes before you upload. One might supply a person or product, while the other supplies a background, another person, or a visual style. The prompt tells the model how those references should become one composition.
The travel example above uses a portrait and a Paris scene. The product example uses a bottle and an interior. Both include the original images, the generated result, and the full prompt, so you can inspect the relationship between the inputs and the output.
Upload one JPG, PNG, or WebP into each slot, up to 10MB per image. Clear source images help the model recognize the details you want to preserve. Select the files locally before signing in; they upload only when you start a merge.
For products and backgrounds, place the main subject in the first slot and the intended scene in the second. For portraits, use one person in each slot. For a room makeover, use the existing room first and the design reference second.
| Style | Use it when | Clarify in the prompt |
|---|
| Natural | A subject belongs in a new scene | Position, scale, and what to preserve |
| People | Two people should share one photo | Who stands where, framing, and setting |
| Product | A product belongs in a background | Surface contact, size, and label details |
| Room | A room should follow a design reference | Architecture to keep and furnishings to change |
| Creative | A subject should adopt a visual style | Main subject, palette, and degree of change |
You do not need to fill in the prompt to use a preset. Add your own instructions when the position, framing, or details matter.
A useful prompt answers three questions: what comes from each image, where it belongs, and what should stay recognizable. Start with one specific instruction, then add the important constraints.
For a person and a destination:
Place the person from image 1 in the riverside scene from image 2. Keep the face, hairstyle, and clothing recognizable. Match the scene's camera height, daylight, and shadows. Leave space around the person for a landscape composition.
For a product and an interior:
Place the bottle from image 1 on the table in image 2. Preserve its shape, amber color, and label placement. Match its size to the furniture, add a contact shadow, and use the same window light as the room.
These are starting prompts you can adapt. The View prompt control on each example shows the exact instructions associated with that displayed result.
Use Auto if you do not have a specific format in mind. Choose 1:1 for square layouts, 4:5 for portrait layouts, or 16:9 for a wide scene. Sign in when you are ready, review the credit cost beside the button, and start the merge.
New accounts receive 3 free 1K generations. Each new result uses one credit. Free-credit downloads include a watermark; paid generations download without one. View pricing for the available plans and one-time packs.
Compare the result with both inputs. Check faces, hands, product labels, straight architectural lines, and shadows where subjects meet the scene. AI generation may reinterpret these details.
If the image needs another attempt, change the instruction that caused the problem. For example, specify “stand to the left” instead of adding more general style adjectives. A clearer input can be more useful than a longer prompt.
The result downloads as a PNG. Your original files are not overwritten. For portrait-specific source selection and troubleshooting, see the AI photo combiner guide.
