# Using ModelVerse API in ComfyUI
## About ComfyUI
ComfyUI is a node-based workflow interface for Stable Diffusion GUI. By breaking down the workflow of Stable Diffusion into nodes, it enables more precise workflow customization and comprehensive process replication.
## Preparation
### Step 1: Install ComfyUI
Please refer to the [official ComfyUI documentation](https://www.comfy.org/en-us/) for installation instructions.
### Step 2: Install SCloud Plugin
1. Open the **Manager** in the top right corner of ComfyUI, and click **Custom Nodes Manager**.
2. Search for "ComfyUI-SCloud".
3. Click **Install**.
4. Restart ComfyUI.
### Step 3: Create Modelverse Node
You can find and add the custom plugin nodes we developed in the node list on the left, or right-click on the canvas and select nodes under the "UCLOUD_MODELVERSE" category.
### Step 4: Obtain UModelVerse API Key
1. [Click here to obtain your API Key](https://console.scloudsg.com/modelverse/experience/api-keys).
2. Add the obtained API Key to the **Modelverse Client** node.
## Quick Start: Auto-Load Workflows
Now, your notebook can easily generate viral pet selfies for social media, like making an orange cat travel to Paris:
**Prompt**:
> This is an iPhone selfie perspective photograph, orange tabby cat wearing sunglasses, sitting in front of Eiffel Tower in Paris, happy expression, warm sunset lighting, travel photography style.
## ModelVerse Core Nodes Introduction
The ModelVerse plugin not only supports standard text-to-image and image-to-image processes but also introduces **text generation** functionalities. Combined with AI prompt optimization, it allows you to build more robust AI workflows in ComfyUI.
- **Modelverse Client**: The entry point for configuring the API Key and the starting point for all workflows.
- **Modelverse Chat**: A text generation node that can be used to generate, optimize, or rewrite prompts.
- **Modelverse [Model Name] Text2Image**: A text-to-image node supporting advanced models like `Flux Kontext Max` and `Flux Kontext Pro`.
- **Modelverse [Model Name] T2V/I2V**: A video generation node, `T2V` supports text-to-video, and `I2V` supports image-to-video, currently supporting the `Wan-AI` series models.
- **Modelverse [Model Name] Edit**: An image editing node for local optimization or style adjustments of generated images, supporting `Step1X Edit`.
## Text-to-Image
With ModelVerse's `Text2Image` node, you can easily turn your textual creativity into high-quality images.
## Text-to-Video
Using the `T2V` node, you can now directly convert text descriptions into dynamic videos within ComfyUI.
## Advanced Workflow: Text Generation -> Image/Video
This is a unique advantage of the ModelVerse plugin. You can first use the `Modelverse Chat` node (such as the `zai-org/glm-4.5` model) to generate or optimize prompts and then directly input the output text into `Text2Image` or `T2V` nodes to achieve fully automated content creation.
This workflow greatly enhances creative efficiency and diversity.
## Batch Processing of Multiple Images
Use **Flux Kontext Pro (Multi-inputs)** to batch generate series of works:
With a single input, you can generate simultaneously:
## Detail Optimization and Image Editing
For unsatisfactory details, use **Step1X-Edit** for fine-tuning:
```
# Brighten eye reflections
Add bright reflection in cat's eyes
# Adjust whisker details
Enhance whiskers detail and texture
# Optimize background blur
Improve background bokeh effect
```
This capability for local editing allows the final result to reach professional photography standards.
## Tips: Practical Prompt Optimization
After extensive testing, we have summarized several key tips:
### 1. Structured Prompting
**Basic Structure**
```
[Camera Device] + [Perspective Description] + [Pet Features] + [Background Environment] + [Lighting and Atmosphere] + [Style Definition]
```
**Practical Example**
```
iPhone 13 Pro selfie perspective +
golden retriever with happy expression +
Santorini blue dome background +
golden hour lighting +
travel photography style
```
### 2. Model Selection Strategy
**Text-to-Image Primary Choice**
- **Flux Kontext Max Text2Image**: Highest quality
**Image Editing Preferred**
- **Flux Kontext Pro**: Strong detail processing
**For Creative Exploration**
- **Flux Dev**: Outstanding stylization ability
**For Batch Production**
- **Multi-inputs version**: Highest efficiency
## Frequently Asked Questions
### Regarding the "import VideoFromFile" Error
If you encounter an "import VideoFromFile Error" after updating our ComfyUI plugin, please upgrade ComfyUI to the latest version and try again.