How to Use Reference Images with Image-to-Image AI
Learn how to add reference photos in WizImage, write a clear transformation prompt, choose a model, and review the result and credit cost.
Image-to-image AI creates a new image using a reference image and a text prompt. In WizImage, you add your own photos in the generator and describe the result you want. A reference can guide the subject, composition, or visual style; it does not guarantee that every original detail will remain unchanged.
1. Add your own reference photos
Open the image generator. Use the + control in the reference image area, or drop an image into that area. WizImage accepts JPEG, PNG, and WebP files up to 10 MB each, with up to five reference images per request.
Choose a clear image that shows the subject you want the model to work with. If you add several references, describe what each one contributes. For example, use one photo for the subject and another for the intended color palette.
An example shown beside a gallery prompt is inspiration, not an automatically uploaded reference. To use a photo as an input, add a file you have permission to use in the reference image area.
2. Describe the change and the details that matter
A useful transformation prompt identifies the subject, the requested change, and the elements you would like to retain. These are instructions to the model, not a promise of exact preservation.
For a product scene, a starting prompt could be:
Use the uploaded bottle as the main subject. Place it on a pale stone surface with soft window light from the left and an uncluttered warm-gray background. Keep the bottle upright and centered. Aim to retain its shape and label layout. Do not add extra objects or new lettering.
This is an editable example prompt, not a measured result or a before-and-after demonstration. After generating, check the shape, cap, label, and lettering against the original photo before using the image.
3. Choose a model and review the cost
The generator shows the available models and the output settings supported by the selected model. Compare the model options and credit costs, then choose the aspect ratio and resolution that suit your task.
Generation requires an account and uses credits. Review the cost on the generation control before submitting; changing the model, resolution, or number of references can affect it. Current subscriptions and credits are described on the pricing page.
4. Review the image before refining it
Check the result for the specific details that matter to your task. For a portrait, inspect the face and expression. For a product, inspect the silhouette, materials, and text. A more specific prompt may help, but cannot guarantee identity, exact wording, or product accuracy.
When refining, change one part of the prompt at a time so you can assess the effect. For example, keep the subject and scene description and change only the lighting. Another generation uses credits, so review your settings first.
When should you start from text instead?
If you do not have an existing image or need a completely new scene, describe the image without adding a reference. Read the text-to-image and image-to-image comparison to choose an approach, or explore image-to-image prompt examples.