bailian-image-gen（白莲图像生成器）

Alibaba Cloud Bailian Image Generation

Based on Alibaba Cloud Bailian platform's qwen-image-2.0 model, providing high-quality text-to-image and image-to-image capabilities.

Features

• Text-to-Image - Generate high-quality images from text descriptions
• Image-to-Image - Generate new images from reference image + text description
• ComfyUI Integration - Complete ComfyUI custom nodes
• Auto Compression - Automatically handle reference image size limits
• Synchronous Calls - No async task waiting, instant results

Use Cases

• C-end App product promotion image generation
• E-commerce product image style transfer and background replacement
• Marketing material batch generation
• Creative image editing

Quick Start

Environment Setup

1. Install dependencies:

   pip install requests Pillow

1. Configure API Key (two methods):

Method 1: .env file (recommended)

   cp .env.example .env
   # Edit .env file
   DASHSCOPE_API_KEY=your-api-key-here

Method 2: Environment variable

   export DASHSCOPE_API_KEY=your-api-key

Get API Key from bailian.console.aliyun.com

Command Line Usage

# Text-to-Image
python scripts/bailian_image_gen.py --mode t2i --prompt "A cute orange cat, high quality" --output cat.png

# Image-to-Image
python scripts/bailian_image_gen.py --mode i2i --prompt "Modern minimalist living room scene, warm lighting" --reference-image product.jpg --output result.png

Python API

from scripts.bailian_image_gen import QwenImageGenerator

# Initialize
client = QwenImageGenerator()

# Text-to-Image
result = client.text_to_image(
    prompt="A cute orange cat, high quality",
    size="1024*1024"
)
url = client.extract_image_url(result)
client.download_image(url, "output.png")

# Image-to-Image
result = client.image_to_image(
    prompt="Modern minimalist living room scene, warm lighting",
    reference_image_path="product.jpg"
)
url = client.extract_image_url(result)
client.download_image(url, "output.png")

ComfyUI Integration

Install Nodes

1. Copy files to ComfyUI:

   cp scripts/bailian_image_gen.py /path/to/ComfyUI/custom_nodes/
   cp scripts/comfyui_bailian_node.py /path/to/ComfyUI/custom_nodes/

1. Configure API Key in ComfyUI directory:

   echo "DASHSCOPE_API_KEY=your-api-key" > /path/to/ComfyUI/.env

1. Restart ComfyUI

Available Nodes

Search "Bailian" in ComfyUI to find these nodes:

BailianText2Image

• Inputs: prompt (STRING), size (COMBO), seed (INT)
• Output: image (IMAGE)

BailianImage2Image

• Inputs: image (IMAGE), prompt (STRING), size (COMBO), seed (INT)
• Output: image (IMAGE)

Workflow Example

Import assets/comfyui_workflow.json for product promotion image generation example.

Typical workflow:

[Load Image] --> [BailianImage2Image] --> [Save Image]
                      ^
                prompt: "Modern minimalist living room, warm lighting"

Parameters

Supported Image Sizes

• 1024*1024 - Square (recommended)
• 1024*768 - Landscape
• 768*1024 - Portrait
• 2048*2048 - High resolution

Prompt Tips

Product Promotion Template

[Product] placed in [Scene], [Style], [Lighting], [Quality requirements]

Examples:

• "Smartwatch placed on minimalist white marble desktop, Nordic minimalist style, natural light, product photography quality"
• "Sneakers placed on wooden floor, city skyline background, fashion magazine style, soft side lighting"
• "Cosmetics placed on dressing table, surrounded by flowers and perfume bottles, luxury style, warm lighting"

File Structure

bailian-image-gen/
├── .env.example              # API Key config example
├── README.md                 # Detailed documentation
├── requirements.txt          # Dependencies
├── SKILL.md                  # This file
├── assets/
│   └── comfyui_workflow.json # ComfyUI workflow example
└── scripts/
    ├── bailian_image_gen.py      # Core script
    └── comfyui_bailian_node.py   # ComfyUI nodes

Notes

1. API Key - Requires Alibaba Cloud Bailian platform account and API Key
2. Image Compression - Reference images are automatically compressed to meet API limits
3. Network Requirements - Requires access to Alibaba Cloud dashscope service
4. Synchronous Calls - qwen-image-2.0 uses synchronous calls, no async task waiting

Error Handling

Common errors and solutions:

References

• [Alibaba Cloud Bailian Console]
• [OpenClaw Documentation]
• [ComfyUI GitHub]

Author

@navygo