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

# Generate Image

> Create or edit an image from a prompt, fully keyboard-driven.

**Generate Image** creates a brand-new image from a description, or edits a
photo you provide — all from a floating panel, fully keyboard-driven.

## Open it

Press <kbd>⌃</kbd> <kbd>⌥</kbd> <kbd>G</kbd> from any app.

## How it works

<Steps>
  <Step title="Pick a model">
    Press <kbd>⌃</kbd> <kbd>⌥</kbd> <kbd>G</kbd>. A row of model tiles floats
    up — pick with <kbd>←</kbd> <kbd>→</kbd> (or <kbd>1</kbd>–<kbd>3</kbd>)
    and confirm with <kbd>↩</kbd>.
  </Step>

  <Step title="Choose a mode">
    <kbd>⇥</kbd> toggles between modes:

    * **Create new** — makes an image from your words.
    * **Edit a photo** — changes a photo you add. Attach up to **10**
      reference photos via **Choose Photo**, **Snip Screen**, <kbd>⌘</kbd>
      <kbd>V</kbd>, or drag-and-drop.
  </Step>

  <Step title="Generate">
    Type your description, set aspect ratio / background if you like, and
    press <kbd>↩</kbd> to generate.
  </Step>

  <Step title="Use the result">
    The result appears **in the panel**, the **hosted link is auto-copied**,
    and the **file auto-saves to your Desktop**. The real per-image cost is
    shown.
  </Step>
</Steps>

## Models

**OpenAI**, **Nano Banana**, **Seedream**.

## Aspect ratios

Auto · Square (1:1) · Wide (16:9) · Tall (9:16) · Landscape (4:3) · Portrait
(3:4) · Social (4:5) · Frame (5:4).

## Background

Auto · Transparent · Opaque.

## History

<kbd>⌘</kbd> <kbd>H</kbd> shows your **last 50 generations** — press
<kbd>↩</kbd> on one to bring it back. The panel remembers where you dragged
it.

## Options

Settings → Tools → Generate Image:

* **When the image is ready, copy** — *the link* or *the image itself* (the
  file is saved to your Desktop either way).
* **Default model**, **aspect ratio**, and **background**.

<Tip>
  Not sure how to phrase a prompt? Start from [Image to
  Prompt](/tools/image-to-prompt) on a reference you like, then adjust the
  wording.
</Tip>

## Billing

Generation runs on **your personal activation token** and is **billed per
person**, with a **monthly quota** of images. If a failure occurs, the
panel's **More details** button shows exactly what the service returned.