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

# Image Generation & Understanding

> Generate, edit, and understand images

Profy Expert comes with a full suite of image capabilities — you can generate images with natural language, edit existing images, or have Expert analyze and understand image content.

## Core Capabilities

<CardGroup cols={2}>
  <Card title="Text-to-Image" icon="wand-magic-sparkles">
    Describe the visual you want in words, and Expert automatically generates the corresponding image.
  </Card>

  <Card title="Image Editing" icon="pen-to-square">
    Upload an image and tell Expert what changes you'd like — swap the background, adjust the tone, remove elements, and more.
  </Card>

  <Card title="Image Understanding" icon="eye">
    Upload an image and ask questions directly. Expert can identify the content and answer related questions.
  </Card>

  <Card title="Multiple Models" icon="sliders">
    Switch between different image generation models during a conversation — each model has its own style and strengths.
  </Card>
</CardGroup>

## How to Use

### Generating Images

Simply describe the image you want in the conversation:

* "Draw a cat wearing a spacesuit with a starry sky background"
* "Generate a minimalist tech company logo in blue tones"
* "Create a watercolor landscape painting with mountains reflected in water"

Expert will choose an appropriate model and generate the image based on your description. You can also specify your preferred image model in the conversation settings.

### Editing Images

<Steps>
  <Step title="Upload an Image">
    Drag and drop the image you want to edit into the chat, or click the attachment button to upload.
  </Step>

  <Step title="Describe Your Changes">
    Tell Expert what modifications you'd like in natural language, such as "change the background to a beach" or "remove the watermark in the bottom right."
  </Step>

  <Step title="Review the Result">
    Expert will return the edited image once processing is complete. If you're not satisfied, you can continue requesting adjustments.
  </Step>
</Steps>

### Image Understanding

After uploading any image, you can ask Expert questions:

* "What's in this image?"
* "Describe the layout of this design mockup for me"
* "What text is in the image?"

Expert will analyze the image and provide a detailed response.

## Use Cases

<AccordionGroup>
  <Accordion title="Design Assets">
    Quickly generate UI prototypes, icons, illustrations, banners, and other design assets — ideal for visual exploration in the early creative stages.
  </Accordion>

  <Accordion title="Photo Editing">
    Swap backgrounds, adjust tones, remove blemishes, extend the canvas — no need to open professional editing software.
  </Accordion>

  <Accordion title="Visual Brainstorming">
    Iterate on image effects through multi-turn conversations, using natural language like "make it brighter" or "switch to a vintage style" until you're satisfied.
  </Accordion>

  <Accordion title="Content Creation">
    Quickly create images for articles, social media, presentations, and more — with support for various styles and dimensions.
  </Accordion>
</AccordionGroup>

## Tips for Better Results

<Tip>
  Good prompts significantly improve generation quality. Try describing your needs along these dimensions:

  * **Subject**: What you want to draw (a cat, a building, a person)
  * **Style**: What artistic style (realistic, cartoon, watercolor, pixel art)
  * **Composition**: Layout of the scene (close-up, panorama, bird's-eye view)
  * **Color Tone**: Main colors and mood (bright, dark, warm tones)
  * **Details**: Important supplementary details (lighting direction, materials, expressions)
</Tip>

<Warning>
  Avoid overly vague descriptions. "Draw a nice picture" is far less effective than "Draw an orange Garfield cat sitting on a windowsill watching the rain, watercolor style."
</Warning>

## Billing

Image generation is billed per call, with each generation consuming a certain number of credits. The exact cost depends on the selected model and parameter configuration — refer to the real-time pricing shown in the product.

Image understanding and analysis follows the conversation model's token-based billing and incurs no additional charges.