Image to Image AI
Image to image AI starts from your picture: restyle a photo, change the scene, or combine up to 3 images into one.
Image to image AI starts from your picture instead of a blank canvas. Use one photo to restyle it or change the scene, or up to 3 images to combine a person, a product and a background into one new image.
How it works
- Upload 1–3 reference images.
- Describe the image you want, for example “turn this photo into a watercolor painting”.
- Choose the aspect ratio and 1K or 2K.
- Download the new image, or press Continue editing to refine it.
FAQ
- What is image to image AI?
- The model takes your image as the starting point and follows your prompt. It keeps the composition, people or products of the original unless you ask it to change them.
- Can I combine several images?
- Yes, up to 3. Refer to them by order in the prompt, for example “put the product from image 1 on the table in image 2”.
- How do I keep parts of the photo the same?
- Say what to keep, for example “keep the face and pose exactly the same”. Short, specific prompts give the most faithful results.
- How many credits does it use?
- Most models use 10 credits per image at 1K. Premium models and 2K use 15 to 50; the menu shows the exact number before you run. With Seedream 5.5, using 2 or 3 reference images costs 15 credits at 1K instead of 10.