Can I combine two photos of people?
Yes, but explain the role of each person and their position in the final scene. If both references are equally important, specify who is in the foreground and how the interaction should look.
AI IMAGE COMPOSER
AI Image Composer for Combining Two Reference Images
Upload two reference images and describe how they should work together. Combine a person, outfit, prop, or scene into one new image with a clearer creative direction.
- Two reference images
- Relationship-based prompt
- One composed output
Compose result
Upload two images and add a prompt to start.
✦
No result yet
Upload two references, add a prompt, then tap Start compose.
History
Generation history is kept for 7 days only. Please download in time.
No generations in the last 7 days. Generated files are kept for 7 days only.
How to combine two images with more control
The strongest compositions give each reference a job. Instead of asking the model to “merge these,” explain what should be preserved from image A, what should come from image B, and how the final scene should connect them.
A practical workflow
- 1
Assign a role to each image
Decide whether image A supplies the subject, pose, or face and whether image B supplies the clothing, object, environment, or lighting. Clear roles reduce ambiguous blending.
- 2
Upload readable references
Use images with a visible main subject and enough resolution to identify the important details. Avoid two busy images when you want one clean focal point.
- 3
Describe the relationship
State where the subjects belong, what they are doing together, and which details must remain recognizable. Mention camera angle, lighting, and background when they matter to the result.
- 4
Review the output and refine one variable
If the result is close, change one instruction at a time: identity, clothing, pose, object placement, or background. Rewriting every detail at once makes it harder to learn which constraint was unclear.
Prompt ingredients that improve composition
A useful prompt names the source of each important detail and describes the final relationship. This example can be adapted for people, products, props, or locations.
Use image A as the person and pose reference. Use image B as the jacket and color reference. Create one natural portrait of the person wearing that jacket in a simple studio setting. Preserve the person's visible features, keep the jacket's main shape and colors, match the lighting, and do not add extra people or unrelated objects.
Common problems and what to change
- Both images compete for attention: name one primary subject and make the second reference provide only the requested detail.
- The clothing or object disappears: describe its position, material, color, and relationship to the subject instead of only naming it.
- Faces or proportions look unstable: use clearer, similarly framed references and shorten the prompt to the highest-priority constraints.
- The scene is almost right but needs cleanup: use the image editor for a targeted correction instead of changing the entire composition.
AI image composer FAQ
Will the output preserve an exact face or logo?
Reference-based generation can drift, especially with small or low-quality source details. Treat the references as guidance, inspect the output, and avoid using generated results where exact identity or trademark accuracy is required without review.
Why does the result look like a collage?
The prompt may not define a shared camera angle, lighting setup, or scene. Add those relationships and state which reference supplies the main subject.
When should I use the image editor instead?
Use the composer when you need to combine two sources. Use the editor when one existing image is already correct and you only need a focused change such as removing an object or adjusting a background.