Image editing in Dreem is designed to feel like post-production for AI-generated content—so you don’t have to start over every time you need a change. Instead of regenerating entire images, you can refine specific details while keeping the original composition and context intact.
Open any image to make focused updates using text prompts (e.g., “change the shoes to loafers”) or the drawing/brush tool to indicate exactly where edits should be applied. Each request generates a new version, allowing you to quickly iterate, compare results, and revert to earlier versions as needed.
Watch the tutorial
Prefer a quick walkthrough? Watch this short tutorial to see how Image Editing works in Dreem, including how to select an image, mark the area you want to edit, describe your change, and compare versions.
You can also follow the step-by-step guide below.
Image Editing
1. Select an Image to Edit
After generating your images, click on the image you’d like to edit directly within the canvas. This will open the image in a larger view, where you can begin making adjustments or requesting changes.
2. Open Editing Tools
Click Edit to access the prompt box and mark-up (brush) tool.
3. Make Your Edits
Use the prompt to describe the changes you’d like to make (e.g., “add a logo” or “change the shoes”). You can make multiple edits at once by selecting a brush color, marking up different areas of the image, and describing each change in the prompt. The brush color links each marked area to its corresponding instruction.
You can also use a reference image to guide your edits. Choose from an originally uploaded image, a previous generation, or upload a new reference. In the prompt, type @ to select and reference an image—available options will appear for you to choose from.
Each request generates a new version while keeping previous versions accessible.
Dreem processes your edit request and generates a new version of the image, reflecting your specified changes.
Useful ways to use Image Editing
Image Editing can be used for more than changing or replacing items. It is also helpful when you want to refine a specific part of an image without regenerating the full result.
Pro-tip: For better result, you should describe clearly what should be removed/adjusted, and what should stay the same. Remember to keep the selected area focused and attach a clear reference image with one main request at a time
Add makeup or bolder styling to models
If you want the model to appear more styled, polished or campaign-ready, you can use Image Editing to add makeup or enhance facial styling.
Refine a specific area of image
You can also use Image Editing when a specific part of the image needs extra refinement. Instead of starting over, select the area you want to improve and describe the adjustment you need.
Clean up unwanted duplicates or extra elements
If your image includes extra elements you do not want such as a duplicated person, partial body, extra object, or background distraction, you can use Image Editing to clean up that specific area.
Version Compare
Once a new image has been generated, you can use Version Compare in the edit view to see a side-by-side comparison of the original and updated versions. Use the slider to move between the two images and closely compare changes.
The latest version (e.g., V2) will be displayed by default, while earlier versions remain accessible.
You can easily navigate between versions by clicking on the image previews and restore a previous version if needed. Using the three-dot icon, you can download images or choose to use the V1 output.
Hovering over the preview of the new version also allows you to view the original edit prompt associated with that generation.
Input Compare
When navigating back to the main view, you can compare the original input images to the generated result.
Click on the image previews to enlarge the originals and more closely assess accuracy.
