What image to image AI does
Image to image starts with a photo you already have. Tell the model what should change and what should stay, such as a bottle’s label, a portrait’s pose or the angle of a chair.
What people use it for
- Product photos: keep the bottle and try a surface such as marble, linen or wet stone.
- Portraits: replace the background, soften shadows or change the season.
- Interiors: try a new palette before buying furniture.
- Sketches: turn a line drawing into an illustration.
- Social images: choose a portrait or landscape frame for the same photo.
Image to Image AI Examples
These image-to-image edits change the background, material or lighting while keeping the same bottle as a reference. Compare the input and edited result, then open the prompt to see the actual instruction. Results can vary.
Input

Edited result

View prompt
Replace only the plain warm grey wall behind the bottle with a deep midnight-blue studio backdrop. Keep the pale oak table, amber glass bottle, black cap and cream label with green circle unchanged. Photorealistic product photography.
Input

Edited result

View prompt
Change the amber glass body of the bottle to matte pale limestone with a natural fine stone texture. Keep the bottle silhouette, black cap, cream label with green circle, oak table, grey background and camera position unchanged. Photorealistic product photography.
Input

Edited result

View prompt
Relight the scene dramatically with a strong golden sunbeam entering from the upper right. Add hard diagonal stripes of light and shadow from window blinds across the grey wall and oak tabletop. Keep the amber bottle, black cap, cream label and green circle unchanged.
The generator guide shows a watercolor-style edit and explains how these examples were made.
How the editor works
- Pick a photo. Upload a JPG, PNG, or WebP file up to 10 MB.
- Describe one change. Try “Replace the background with a rainy Tokyo street. Keep the face and denim jacket.”
- Pick a shape. Match the original or choose a different aspect ratio.
- Download the result. The download link appears below the editor when the image is ready.
What makes a good result
- Name what to keep. If the label or the pose matters, say so in the prompt.
- Describe the scene. “A wall of green plants behind her” is more specific than “make it look better.”
- Change one thing at a time. If the light and background both need work, start with whichever matters more.
What the editor can miss
- Text in the photo. Small text on packaging can change. Check labels and key details before using the output.
- Faces. The model can change features, especially if the face is turned away or in shadow.
- Pixel-perfect edits. The model rebuilds parts of the photo, so fine details can shift.
- Groups. If one person matters most, try cropping closer to them first.
Free access and credits
Free generation is subject to availability. Sign in with Google to upload a photo. Paid generations use credits, and the cost can vary by model. See free access or pricing for details.
Where to go next
- The AI image to image generator has more prompt examples.
- Guest access explains what you can do without an account.
- Pricing lists credit packs and subscriptions.
Frequently asked questions
Is this free to use?
Yes, when free capacity is available. Sign in with Google to upload a photo. Paid generations use credits; the cost depends on the model.
What is the difference between image to image and text to image?
Text to image starts with a written description. Image to image starts with a photo and uses your description to guide what changes.
Which file types and sizes does the editor accept?
JPG, PNG and WebP files up to 10 MB.
Do I keep the rights to the images I generate?
You keep the rights to your original upload. Generated output may be used commercially, subject to the terms.
Does it work on a phone?
Yes. You can pick a photo from your camera roll and use the editor in your browser.
Which model does the editor use?
Google's Nano Banana Edit model through Kie.ai.