Upload the source
Choose a clear portrait, product image, illustration, or composition that contains the visual information you want to keep.
Upload a reference image, explain what should change, and generate a new version while preserving the visual details that matter to your idea or brand.
Specify what to replace, preserve, relight, restyle, or recompose in the uploaded image.
Image-to-image AI is strongest when the upload establishes identity or composition and the prompt clearly separates preserved details from requested changes.
Choose a clear portrait, product image, illustration, or composition that contains the visual information you want to keep.
State what must remain unchanged, then describe the new background, lighting, wardrobe, material, camera treatment, or style.
Keep successful constraints, remove ambiguous language, and regenerate when identity, geometry, or small details drift.
Creators often need a controlled variation rather than a completely new image. Reference-led editing makes the starting point visible to the model.
Change the mood, wardrobe, backdrop, lens treatment, or editorial direction while asking the model to retain identity.
Move a product or subject into a cleaner studio, lifestyle setting, campaign environment, or presentation-ready scene.
Explore different materials, colors, surfaces, props, and art directions from an existing product reference.
Adapt a visual for square, landscape, or portrait placement and direct where negative space should appear.
Simple restyles benefit from fast iteration. Complex transformations, text, or fine product details often justify a model with stronger instruction following.
A practical default for prompt-led images, quick variations, and responsive editing.
Choose it for intricate concepts, typography-aware compositions, and commercial polish.
Use the premium Gemini tier when realism, finish, and careful visual structure matter most.
A lower-credit option for exploring ideas before moving to a higher-fidelity final render.
A strong AI photo editing prompt contains a preservation clause, the desired transformation, and the intended visual finish.
Browse more prompt ideasKeep the bottle shape, label, cap, and camera angle unchanged. Replace the background with pale travertine, warm sunrise light, soft natural shadow, premium skincare campaign.
Preserve the person’s facial identity and expression. Change the wardrobe to a structured black blazer, add soft Rembrandt light, neutral gray editorial studio, realistic skin texture.
Recompose this image for vertical 4:5. Keep the subject scale natural, extend the environment above and below, and create clean negative space on the upper left for campaign copy.
An AI photo editor uses a source image plus written instructions to create a transformed version. It can change backgrounds, lighting, styling, composition, materials, or mood without manual masking.
Text-to-image begins from a description alone. Image-to-image also receives a visual reference, which can provide identity, product shape, pose, palette, or composition.
Use a clear reference and explicitly list the features that must not change. Consistency is model-dependent, so complex identity or geometry may require several controlled iterations.
Yes. Describe the replacement setting, light direction, surface, shadows, depth of field, and the elements that should remain untouched.
Yes. You can preserve the product and generate new campaign environments, props, layouts, and aspect ratios. Always review labels and exact product geometry before publishing.
The workflow generates a new output from your reference and prompt. Your source remains the reference rather than being overwritten in place.
Upload a reference, define what stays, and describe the transformation in plain language.
Edit an image with AI