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

# Generate & Edit Images with ImageZen

> ImageZen is Artnex's AI image generation and editing hub. Choose from 14 models, write a prompt, and create stunning visuals in seconds.

ImageZen is the core image creation tool in Artnex. It brings together text-to-image generation, image-to-image editing, prompt enhancement, and image-to-prompt extraction — all in one interface. Whether you're creating from scratch or transforming an existing photo, ImageZen gives you a full model lineup to match your quality and speed needs.

## Navigate to ImageZen

Go to **Tools > Image > ImageZen** in the Artnex sidebar.

## Available models

ImageZen groups its models into three tiers: Fast, Pro, and Edit. Select a model by clicking the model button in the bottom control bar, then choosing from the dialog.

<Tabs>
  <Tab title="Fast models">
    Fast models prioritize speed and are ideal for iteration and concept exploration.

    | Model         | Provider  | Credits per generation | Best for                          |
    | ------------- | --------- | ---------------------- | --------------------------------- |
    | Z-Image Turbo | Wavespeed | 100                    | Rapid concept testing             |
    | Nano Banana   | Google    | 240                    | Fast generation with good quality |

    <Note>Z-Image Turbo supports a **seed** parameter so you can reproduce or vary a specific result. Use `-1` for a random seed.</Note>
  </Tab>

  <Tab title="Pro models">
    Pro models produce higher quality output and offer more resolution and control options.

    | Model                 | Provider  | Credits per generation | Notes                                               |
    | --------------------- | --------- | ---------------------- | --------------------------------------------------- |
    | Seedream v4           | Wavespeed | 200                    | High quality with strong prompt adherence           |
    | Flux 2 Pro            | Wavespeed | 250                    | Flexible resolutions up to 1536 px                  |
    | Nano Banana Pro       | Google    | 800                    | Multi-image generation; 1K–4K resolution            |
    | Seedream v4.5         | Bytedance | 800                    | 4K output with crisp typography rendering           |
    | Nano Banana Pro Ultra | Google    | 1,000                  | 8K ultra-high resolution; 1K–8K resolution selector |
    | GPT Image 2           | OpenAI    | 60–3,960\*             | Photorealistic; 1K–4K with quality control          |

    \*GPT Image 2 uses dynamic pricing based on resolution (1K/2K/4K) and quality (low/medium/high).

    <Tip>If you need sharp text inside your image — labels, signs, logos — **Seedream v4.5** handles typography better than most models.</Tip>
  </Tab>

  <Tab title="Edit models">
    Edit models require one or more reference images. They modify or blend those images according to your prompt.

    | Model                      | Provider  | Credits per edit | Notes                                       |
    | -------------------------- | --------- | ---------------- | ------------------------------------------- |
    | Nano Banana Edit           | Google    | 240              | Fast editing; supports aspect ratio control |
    | Seedream v4 Edit           | Bytedance | 200              | Creative transformations up to 4096×4096    |
    | Nano Banana Pro Edit       | Google    | 800              | Professional editing; 1K–4K resolution      |
    | Seedream v4.5 Edit         | Bytedance | 800              | Face and lighting preservation; 4K output   |
    | Nano Banana Pro Edit Ultra | Google    | 1,000            | Ultra-high quality 4K/8K editing            |
    | GPT Image 2 Edit           | OpenAI    | 180–3,960\*      | Blend up to 10 reference images; 1K–4K      |

    \*GPT Image 2 Edit uses the same dynamic pricing grid as GPT Image 2.

    <Note>When you select an Edit model, the interface switches to edit mode automatically and requires at least one uploaded reference image before you can generate.</Note>
  </Tab>
</Tabs>

## Generate an image

<Steps>
  <Step title="Open ImageZen">
    Navigate to **Tools > Image > ImageZen**.
  </Step>

  <Step title="Write a prompt">
    Type your description in the prompt field at the bottom of the screen. Be specific: include subject, style, lighting, and mood.
  </Step>

  <Step title="Select a model">
    Click the model button in the control bar to open the model picker. Choose the model that fits your quality and credit budget.
  </Step>

  <Step title="Set aspect ratio">
    Use the aspect ratio selector next to the model button. Options vary by model — most support 1:1, 16:9, 9:16, and more.
  </Step>

  <Step title="Click Create">
    Click **Create** (with the credit cost displayed) to start generation. A progress indicator appears while your image is being generated.
  </Step>

  <Step title="Download or save">
    When the image appears in your creations gallery, hover over it to reveal the download button. All generated images are automatically saved to your Creations library.
  </Step>
</Steps>

## Use image-to-image (edit mode)

Edit models let you upload a reference image and transform it using a text prompt.

<Steps>
  <Step title="Select an Edit model">
    Open the model picker and choose any model from the **Edit Models** group. The interface updates to edit mode.
  </Step>

  <Step title="Upload reference images">
    Click **Upload** in the control bar and select up to 10 images (PNG, JPEG, or WebP). Thumbnails appear in the prompt bar.
  </Step>

  <Step title="Describe the edit">
    Write a prompt describing the change you want — for example, "change the background to a snowy forest" or "make the lighting more dramatic."
  </Step>

  <Step title="Click Edit">
    Click **Edit** to run the generation. The edited image is saved to your Creations gallery automatically.
  </Step>
</Steps>

<Warning>Edit models require at least one uploaded image. The **Edit** button stays disabled until you upload a reference.</Warning>

## Prompt Enhancer

The Prompt Enhancer rewrites your prompt to be more detailed and model-optimized, without changing your core idea.

To use it in ImageZen:

1. Type your prompt in the prompt field.
2. Click the **sparkle icon** (✦) at the bottom-right of the prompt field.
3. Artnex replaces your prompt with the enhanced version automatically.

You can also use the standalone **Prompt Enhancer** tool at **Tools > Prompt Enhancer** for more control, including context, detail level, and tone settings.

## Image-to-Prompt

Image-to-Prompt analyzes an uploaded image and generates a descriptive prompt you can use — or modify — for generation.

<Steps>
  <Step title="Open the Image-to-Prompt dialog">
    Click the **image icon** at the bottom-right of the prompt field to open the Image-to-Prompt dialog.
  </Step>

  <Step title="Upload an image">
    Drop or select any image file. Artnex analyzes it and returns a detailed prompt describing the scene.
  </Step>

  <Step title="Apply or edit the prompt">
    Click **Apply** to load the generated prompt into the prompt field, then edit it as needed before generating.
  </Step>
</Steps>

## Downloading and saving creations

All generated images are saved automatically to your Creations gallery, accessible from the ImageZen canvas or the global **Creations** section.

To download an image:

* Hover over any image in your gallery and click the **download icon**.
* The file downloads in the format you selected at generation time (JPEG, PNG, or WebP depending on the model).

## Tips for better prompts

<CardGroup cols={2}>
  <Card title="Be specific about style" icon="palette">
    Instead of "a woman," write "a close-up portrait of a woman in soft natural light, editorial style, shallow depth of field." Specific style cues produce more consistent results.
  </Card>

  <Card title="Describe lighting" icon="sun">
    Lighting has one of the biggest impacts on image quality. Include terms like "soft studio lighting," "golden hour," "dramatic side lighting," or "overcast diffused light."
  </Card>

  <Card title="Use the Prompt Enhancer" icon="sparkles">
    If you're unsure how to phrase your idea, type a short description and click the sparkle button. The enhancer adds relevant technical details for you.
  </Card>

  <Card title="Iterate with seeds" icon="refresh-cw">
    When using Z-Image Turbo, set a seed to lock in a composition you like, then vary only the prompt. Use -1 to get a new random result each time.
  </Card>
</CardGroup>

<Tip>For portrait or social media content, 4:5 is often a better fit than 1:1 — it fills more of a phone screen and matches Instagram's recommended format.</Tip>