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

# Image generation

# Image Generation

ShuYou supports invoking image generation models via the Vertex AI protocol. This guide explains how to use ShuYou to generate images and save them locally.

<Tip title="💡 About Banana Models">
  Banana is a series of image generation models from Google that can produce high-quality images from text prompts. You can use these models in ShuYou through the Vertex AI protocol.
</Tip>

## Supported Models

The currently supported image generation models include (continuously updated):

* `gemini-3.1-flash-image-preview` (Nano Banana 2)
* `gemini-3-pro-image-preview` (Nano Banana Pro)
* `gemini-2.5-flash-image` (Nano Banana)

<Tip title="📚 More Models">
  Visit the [ShuYou model catalog](https://shuyou.ai/models) to search and view all available image generation models.
</Tip>

## API reference

Image generation APIs with interactive playgrounds:

* [GPT Image 2](/en/api-reference/image-series/openai/gpt-image-2-generate) — `POST /v1/predictions`
* [Gemini 3.1 Flash Image Preview (Nano Banana 2)](/en/api-reference/image-series/gemini/gemini-3.1-flash-image-preview-generate) — `POST /v1/predictions`
* [Gemini 3 Pro Image Preview (Nano Banana Pro)](/en/api-reference/image-series/gemini/gemini-3-pro-image-preview-generate) — `POST /v1/predictions`
* [Gemini 2.5 Flash Image (Nano Banana)](/en/api-reference/image-series/gemini/gemini-2.5-flash-image-generate) — `POST /v1/predictions`

## Reference Documentation

This guide only covers basic usage. For detailed configuration and advanced usage, refer to the official documentation below:

* [Vertex AI Official Documentation](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/model-reference/inference)
* [Vertex AI Nano-Banana Notebook](https://github.com/GoogleCloudPlatform/generative-ai/tree/main/gemini/nano-banana)

## Usage

```Python title="Python" theme={null}
from google import genai
from google.genai import types

client = genai.Client(
    api_key="$ShuYou_API_KEY",  # Replace with your API key
    vertexai=True,
    http_options=types.HttpOptions(
        api_version='v1',
        base_url='https://api.shuyou.ai'
    ),
)

# Streaming call: generate_content_stream
# Non-streaming call: generate_content
prompt = "Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme"

response = client.models.generate_content(
    model="google/gemini-3-pro-image-preview",
    contents=[prompt],
    config=types.GenerateContentConfig(
        response_modalities=["TEXT", "IMAGE"]
    )
)

# Handle text and image responses
for part in response.parts:
    if part.text is not None:
        print(part.text)
    elif part.inline_data is not None:
        # Save the generated image
        image = part.as_image()
        image.save("generated_image.png")
        print("Image saved as generated_image.png")
```

```ts title="TypeScript" theme={null}
const genai = require("@google/genai");

const client = new genai.GoogleGenAI({
  apiKey: "$ShuYou_API_KEY",  // Replace with your API key
  vertexai: true,
  httpOptions: {
    baseUrl: "https://api.shuyou.ai",
    apiVersion: "v1"
  }
});

// Streaming call: generateContentStream
// Non-streaming call: generateContent
const response = await client.models.generateContent({
  model: "google/gemini-3-pro-image-preview",
  contents: "Generate an image of the Eiffel tower with fireworks in the background",
  config: {
    responseModalities: ["TEXT", "IMAGE"],  // Response modalities must be specified
    // For more configuration options, refer to the Vertex AI official documentation
  }
});

console.log(response);
```

## Configuration

### Required Parameters

* **api\_key**: Your ShuYou API key
* **vertexai**: Must be set to `true` to enable the Vertex AI protocol
* **base\_url**: ShuYou Vertex AI endpoint `https://api.shuyou.ai`
* **responseModalities**: Response modalities; image generation must include `["TEXT", "IMAGE"]`

### Invocation Modes

ShuYou supports two invocation modes:

* **Streaming** (`generate_content_stream` / `generateContentStream`): Ideal for scenarios requiring real-time feedback
* **Non-streaming** (`generate_content` / `generateContent`): Returns the complete response at once after processing

<Warning title="⚠️ Response Handling">
  Responses from image generation models may contain both text and images. Iterate over `response.parts` to process all content parts.
</Warning>

## Best Practices

1. Prompt Engineering: Use clear and specific descriptions to achieve better generation quality.
2. Error Handling: Add exception handling to manage potential API call failures.
3. Image Saving: The Python SDK provides a convenient `as_image()` method to convert a response part into a PIL Image object.
4. Model Selection: Choose the appropriate model based on your needs; free models are suitable for testing, while paid models provide higher quality.