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

# AI image generation

> Generate new images from a prompt, or remix existing ones with AI

When your library doesn't have the perfect image, Denada can create one. You can generate brand-new images from a text prompt, or remix an existing image with targeted changes — all without leaving your workflow.

## Generating images from a prompt

Look for **Auto-generate image** wherever you work with images:

* **In the editor** — click any image in the preview and choose **Auto-generate image** from the image menu. See [using images](/editor/using-images).
* **In the image library** — open a collection's menu and choose **Auto-generate image**. The result is saved as a new image in that collection, complete with generated appearance metadata.
* **In chat** — ask Denada for imagery while building a project, and it can generate images as part of the conversation (when your team has enabled it — see below). See the [chat interface](/creating/chat-interface).

The generation dialog lets you write a prompt (e.g., "Students walking on a college campus"), pick a model, and choose an aspect: **Square**, **Landscape**, or **Portrait**.

## Remixing an existing image

Remixing starts from an image you already have and applies changes you describe — "add a gray border", "make the background warmer", "remove the text". It's often faster than regenerating from scratch, and it preserves what already works.

Choose **Remix image** from the image menu to open the remix dialog:

1. Describe how you want to change the image in the prompt field.
2. Optionally drag in additional reference images.
3. Pick a model and an aspect — **Match aspect** keeps the original's proportions, or force **Square**, **Landscape**, or **Portrait**.
4. Send the prompt, then use the eye toggle to flip between the original and the new version.
5. Click **Save** to keep the result, or refine your prompt and try again.

## Where generated images live

Images that Denada generates are collected in the **Generated** collection in your image library, so nothing is lost even if you generated it deep inside a project. Images you generate directly in a collection are saved to that collection.

The **Generated** collection is managed automatically and isn't available to Denada in chat — if you want Denada to reuse a generated image in future projects, move it into one of your published collections. See the [image library](/image-library/collections).

## Enabling AI generation for chat and agents

Team managers control whether Denada and connected AI agents may generate images on their own. In your team's [settings](/team/workspace-settings), the **AI image generation** toggle:

* **Off (the default)** — Denada won't generate images during chat, and connected AI agents can't generate them either. Projects draw on your image library instead.
* **On** — chat and connected agents can generate new images when a project calls for one.

<Note>
  This setting only gates automatic generation. Team members can always generate images manually with **Auto-generate image** and **Remix image**.
</Note>