> ## Documentation Index
> Fetch the complete documentation index at: https://docs.covver.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI image editing

> Generate photos, edit an existing image, merge photos and remove backgrounds with AI in the Builder.

The Builder has two places for AI image work: one-click actions in the floating toolbar, and requests you write to the [Covver Assistant](https://docs.covver.com/ai/overview.md). This page covers both. For cropping, filters and fit, see [images](https://docs.covver.com/builder/images.md).

## Quick actions in the toolbar

Select a single image on the canvas. In the floating toolbar, open **AI** to see the **Image actions** menu.

| Action | What it does |
| - | - |
| **Remove background** | Cuts out the subject and replaces the image with a transparent version. |
| **Replace with AI image** | Swaps the image for a new AI-generated one that fits the surrounding layout. |
| **Generate caption** | Adds a short caption below the image in the tone of the page. |
| **Suggest better crop** | Adjusts the image's shape and crop to improve the composition. |
| **Match page palette** | Adjusts the image's color treatment to harmonize with the rest of the page. |

While an action runs, the image shows a **Processing…** overlay. Only one action runs at a time. If you do not like the result, undo it.

> [!NOTE]
>
> **Remove background** works on images that have finished uploading. If you just pasted or dropped an image, wait a moment for the upload to complete and try again.

## Edit images by asking the assistant

Open the Covver Assistant, select the image you mean, and describe the change. Selecting first matters: the selected image appears as an **Editing:** chip and the assistant knows which photo you are talking about.

> [accordion] Change an existing photo
>
> Ask for an edit that keeps the same photo and changes only what you mention.
>
> * "Same photo, but at night."
> * "Make it black and white."
> * "Give her a red jacket."
>
> The pose, composition and background stay; only the requested detail changes.

> [accordion] Replace a photo with a new one
>
> Ask for a completely different image in the same spot.
>
> * "Replace this with a sunset over the sea."
> * "Swap this for a photo of a vinyl record on a wooden table."

> [accordion] Merge several photos
>
> Select two or more images and ask the assistant to combine them.
>
> * "Merge these photos."
> * "Put the cat from the first photo into the second one."
>
> The combined result replaces one of the selected images.

> [accordion] Generate photos for a new page
>
> When you ask for a full page or a whole issue, the assistant creates the photos the layout needs. Describe the subject and mood in your request ("a moody portrait of a jazz drummer, warm stage light") to steer them.

An image edit typically takes several seconds. While a photo is being created, its frame shows **Generating image…** over a grey placeholder, and the rest of the page is already editable.

## Text in front of or behind a person

Some cover designs place the magazine title behind the subject, so the person appears to stand in front of the lettering. When the assistant composes that style of cover, it cuts the subject out for you. To build the effect by hand, duplicate the photo, run **Remove background** on the copy, and place the cutout above the title in the layer order.

## Use your own photos instead

AI images are optional. Ask the assistant to "use my photos" and it lays out images from your library, or drag any image from your library onto a placeholder to replace it. See [images](https://docs.covver.com/builder/images.md) and [media](https://docs.covver.com/builder/media.md).

## Good to know

* **Generated photos avoid lettering.** The assistant asks for images without text in them, because headlines belong in editable text elements. Add titles and captions as text.
* **There is a per-request cap.** One request generates up to six AI images. Extra photo slots stay as placeholders; ask again for those pages or fill them yourself.
* **Results vary.** Asking again produces a different image. Faces, hands and small details can come out wrong, so check every generated photo before you publish.
* **You are responsible for what you publish.** Review generated images for accuracy and make sure you have the rights to any photos you upload as a starting point.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.