A collection of essential nodes designed to seamlessly integrate Alibaba's Qwen-Image model and its specific LoRA format into the ComfyUI workflow.
A specialized LoRA loader that correctly handles the key conversion for Qwen-Image LoRAs. It features an auto-detect mechanism for the alpha value.

A prompt styler node that quickly appends professional-grade keywords to your prompt, allowing you to easily switch between various artistic styles like cinematic, photorealistic, anime, etc.

Provides a set of recommended aspect ratios for Qwen-Image, along with a "Custom" option for manual input. Ensures output dimensions are valid.

- 启动 ComfyUI。
- 打开 ComfyUI Manager 菜单。
- 点击 "Install Custom Nodes"。
- 搜索
Qwen-Image-Toolkit。 - 点击 "Install" 按钮,安装完成后重启 ComfyUI。
- 进入 ComfyUI 的自定义节点目录:
cd ComfyUI/custom_nodes/ - 克隆本仓库:
git clone https://github.com/Cx330v/Qwen-Image-Toolkit.git
- 重启 ComfyUI。
2025-08-14更新适配里布训练的lora模型
- Qwen-Image LoRA 加载器:
- 将
lora_alpha设为0(默认值) 来自动从adapter_config.json文件中读取 alpha 值。 - 如果需要手动覆盖,可以直接输入一个大于0的 alpha 值。
- 将
- Qwen-Image 提示词:
- 在
text框中输入您的基础提示词。 - 从
style下拉菜单中选择一个风格,对应的风格化关键词会自动附加到您的提示词末尾。
- 在
- Qwen-Image 图像比例:
- 从
aspect_ratio下拉菜单中选择一个预设的比例。 - 或者,选择
Custom 自定义并手动设置custom_width和custom_height。
- 从
This project is licensed under the MIT License. See the LICENSE file for details.
