> ## 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.

# Covver Assistant

> Use the AI assistant inside the Builder to design pages, build whole issues and edit what is already on the page.

Covver Assistant is the AI design assistant built into the Builder. You describe what you want in plain language, and it adds or changes real, editable elements on your pages. Everything it makes lands on the same canvas you edit by hand, so you can keep adjusting it afterwards and undo anything you do not like.

## Open the assistant

> [tab] Desktop
>
> Click **AI** in the left rail of the Builder. The **Covver Assistant** panel slides in next to the canvas. Click **AI** again, or the close button in the panel, to hide it.
>
> You can also press `Cmd+K` (Mac) or `Ctrl+K` (Windows) to open a quick prompt box with suggestions such as **Generate a magazine cover** or **Improve current page**.

> [tab] Phone
>
> Tap **Agent** in the bottom bar. The assistant opens as a sheet over the canvas.

Type your request and press Enter. While the assistant works, the panel shows its progress. When it finishes, it tells you how many elements it added, updated or deleted.

## What it can do

| You ask for | What happens |
| - | - |
| A full page (a cover, a table of contents, an article, a photo page) | The assistant writes the content and a layout engine arranges it on the page so text and photos do not overlap or fall off the edge. |
| A page from a template | It fills one of Covver's curated [templates](https://docs.covver.com/builder/templates.md) with your text and images. |
| A two-page spread | It composes a layout across two facing pages. See [spreads](https://docs.covver.com/builder/spreads.md). |
| A whole issue | Describe the magazine and how many pages you want. The assistant plans every page, then writes and designs them. |
| A change to the current page | Rewrite a text, move or resize things, change colors and fonts, delete elements, or redesign the page while keeping its content. |
| A consistent look | It can set the palette and fonts for the whole magazine so new pages match. |
| A new or edited image | It generates photos, edits an existing photo, or merges several. See [AI image editing](https://docs.covver.com/ai/image-editing.md). |
| Interactive content | Paste a YouTube, Spotify or map link to add an [embed](https://docs.covver.com/builder/embeds.md), ask for a game or carousel [widget](https://docs.covver.com/builder/widgets.md), or describe a custom widget. |
| Your own media | It can place photos, videos and audio from your library on a page. |
| A locked page | It can put a page behind a password, email, phone, quiz or payment. See [page lock](https://docs.covver.com/builder/page-lock.md). |

### Plans and cover options

When you ask for a new page on an empty page, the assistant usually replies with a short plan first. Answer to confirm, change or cancel it. For covers, it can show a few design directions as thumbnails; click one to apply it.

On a magazine with little content, the empty chat offers **Plan the whole magazine**, which starts a full-issue request for you.

### Work on a selection

Select one or more elements on the canvas before you write. They appear above the input as **Editing:** chips, and the assistant treats them as the target of requests like "make this bigger" or "translate this".

### Use a reference image

Click the paperclip, paste, or drop an image into the panel (PNG, JPEG or WebP, up to 5 MB). Then tell the assistant what to do with it, or pick a shortcut:

* **Extract layers** recreates the image as editable text, shapes and image placeholders.
* **Use as background** places the image as a full-page background and extracts the text on top.

### Quick actions on the canvas

Select a single text or image and open **AI** in the floating toolbar for one-click actions such as **Improve writing**, **Make shorter**, **Translate to Spanish**, **Remove background** or **Generate caption**. These run without opening the chat.

## Conversations

Each magazine keeps its own chat history. Click the history button in the panel header to reopen a past conversation or start a **New chat**. When you return to a magazine, the assistant offers to **Continue** the previous conversation.

## What it cannot do

* **It does not publish.** The assistant only edits your draft. You [publish](https://docs.covver.com/builder/publish.md) when you are ready.
* **It does not create video or audio.** It can place videos and audio you already uploaded, but it does not generate them.
* **It works on one request at a time.** Wait for the current request to finish before sending the next one.
* **It caps AI photos per request.** A single request generates a limited number of AI images (up to six). Any extra photo slots stay as placeholders that you can fill yourself or ask for again.
* **It can get things wrong.** Generated text and images are a first draft. Read the copy, check names and facts, and adjust the layout by hand where needed.

> [!TIP]
>
> Not happy with a result? Undo it like any other edit, or ask for "another option" to get a different design with the same content.

## Next steps

> [card] [Prompting tips](https://docs.covver.com/ai/prompting-tips.md)
>
> Get better pages with clearer requests.

> [card] [AI image editing](https://docs.covver.com/ai/image-editing.md)
>
> Generate, edit and cut out photos.

> [card] [Start with AI](https://docs.covver.com/ai/start-with-ai.md)
>
> Build a new magazine from a single description.

> [card] [Covver Builder connector](https://docs.covver.com/mcp/overview.md)
>
> Create magazines from Claude or ChatGPT.


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