Overview
Images created with AI media generation can be edited before they are saved. This allows you to refine a generated image without starting over from a new prompt.
For example, you can change the color of an object, adjust a person's appearance, or describe another specific change that you want to apply to the image.
Editing an AI-Generated Image
- Open the AI media generation flow.
- Generate an image.
- On the image you want to update, click Edit.
- In the edit dialog, describe the change you want to make.
- Click Apply changes.
The updated image is generated and added to the current set of results.
Using Prompt Suggestions
The edit dialog includes suggestion chips for common editing requests. Selecting a suggestion adds a prompt template that you can customize.
For example, if you select a suggestion for changing color, the prompt template helps you specify which object should change and what the new color should be.
You can use a suggestion as-is, edit the suggested prompt, or write your own prompt from scratch.
Reviewing Edited Versions
After an edit is generated, you can continue reviewing the original images and any edited versions created during the same generation session.
You can scroll through the results, choose a specific image to save, or continue editing from a specific version.
Note: Only images that you explicitly save are kept. If you leave the generation flow without saving an image, unsaved generated images are lost.
Credit Usage
Applying edits to a generated image uses credits. The cost is displayed before the edit is applied.
For image generation, editing an image costs the same as generating a single image.
When editing first frames in the AI video generation flow, credits are not charged for the first-frame image edit. However, first-frame images from that flow cannot be saved as standalone image assets.
Limitations
- Image editing is currently available only within the AI media generation flow.
- Saved media library assets cannot currently be edited with this AI image editing flow.
- Background removal and transparent backgrounds are not currently supported.
- Multiple generated images cannot currently be selected and saved together. Each image must be saved individually.