# moark-comfyui

**Repository Path**: moark/moark-comfyui

## Basic Information

- **Project Name**: moark-comfyui
- **Description**: Moark-ComfyUI 是基于 ComfyUI 的自定义节点集合，提供对 Moark AI API 的便捷接入，支持 Qwen 图像生成与 FLUX.1-dev 文生图，并支持通过 URL 加载 LoRA（最多 3 个）。
- **Primary Language**: Python
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 9
- **Forks**: 2
- **Created**: 2025-09-18
- **Last Updated**: 2026-04-10

## Categories & Tags

**Categories**: ai

**Tags**: None

## README

# moark-comfyui

#### Description
Moark-ComfyUI is a set of custom nodes for ComfyUI that integrates with the Moark AI API. It supports text-to-image generation with Qwen Image and FLUX.1-dev, and allows applying up to 3 LoRAs via public URLs.

#### Available Nodes
- Moark API Client: Creates and provides a Moark API client (input: `api_key`).
- Moark Loras: Configures a list of LoRAs (`path` and `weight/scale`).
- Moark Qwen Image Lora: Generates images with the Qwen model, with LoRA support.
- Moark Flux Dev Lora: Generates images with the FLUX.1-dev model, with LoRA support.

#### Installation
1. Navigate to the ComfyUI/custom_nodes directory.
2. Clone this repository: `git clone https://gitee.com/gitee-ai/moark-comfyui.git`
3. Install dependencies:

- Windows (ComfyUI portable): `python -m pip install -r requirements.txt`
- Linux or macOS: `pip install -r requirements.txt`

4. Optional: If you prefer not to expose your API key in the node, rename `config.ini.tmp` to `config.ini` and put your API key inside.
5. Start ComfyUI and use the Moark nodes.

#### Usage
1. Add and connect the following nodes in your workflow:
   - `Moark API Client` → outputs `client`.
   - Optional: `Moark Loras` → outputs `loras` (up to 3).
   - Choose a generation node: `Moark Qwen Image Lora` or `Moark Flux Dev Lora`.
2. Set `prompt`, `negative_prompt`, size, steps, seed, etc. in the generation node.
3. Run the workflow. The node outputs an image tensor, which can be connected to preview or save nodes.

Workflow example (Qwen-Image):
![](./examples/Moark-ComfyUI-Qwen-Image-Example.jpg)

#### How to Apply LoRA (Moark)
1. Since the service runs on the Moark API, local LoRA files are not supported. We support loading LoRA via a publicly accessible URL.
2. Use the `Moark Loras` node to configure LoRAs by providing:
   - `lora_path`: A LoRA URL (e.g., `https://gitee.com/gitee-ai/moark-assets/raw/master/flux.1-dev%2Ftost-2024-10-05-16-42-00-5t1cker-5ty1e.safetensors`).
   - `lora_weight`: Weight in range 0.0 ~ 4.0. Recommended 0.5 ~ 1.5.
3. You can add multiple pairs (up to 3 LoRAs).
4. The URL does not need to be on Hugging Face; any public URL is fine.

#### Contribution
1. Fork this repository
2. Create a new branch (e.g., `feat_xxx`)
3. Commit your changes
4. Open a Pull Request

#### License
This project follows the licenses of its dependencies and upstream projects. Please open an issue if you have questions.