Overview
This node generates a 3D model from 2 to 4 reference images of an object or character. You provide images from different angles (front, left, back, right), and the node creates a 3D mesh in GLB format. The front view is required, and you can optionally add any combination of the other three views for better results.Inputs
| Parameter | Description | Data Type | Required | Range |
|---|---|---|---|---|
image | Front view (0°). Required. | IMAGE | Yes | - |
image_left | Left view (90°), i.e. the subject’s left side. | IMAGE | No | - |
image_back | Back view (180°). | IMAGE | No | - |
image_right | Right view (270°), i.e. the subject’s right side. | IMAGE | No | - |
output_mode | The output mode for the generated model. "geometry" produces a raw mesh, "textured" adds a standard texture, and "detailed" creates a high-detail textured model (default: "textured"). | COMBO | Yes | "geometry""textured""detailed" |
face_limit | Maximum number of faces for the output mesh. Set to -1 for no limit (default: -1). | INT | No | -1 to 100000 |
model_seed | Seed for reproducible model generation. Set to 0 for random (default: 0). | INT | No | 0 to 2147483647 |
auto_size | Automatically size the model to fit within a standard bounding box (default: False). | BOOLEAN | No | True / False |
export_uv | Export UV coordinates with the model (default: True). | BOOLEAN | No | True / False |
compress_geometry | Compress the geometry data to reduce file size (default: False). | BOOLEAN | No | True / False |
image) plus at least one of the other views (image_left, image_back, or image_right). If fewer than 2 images are provided, the node will raise an error.
Outputs
| Output Name | Description | Data Type |
|---|---|---|
model_file | The filename of the generated GLB model (for backward compatibility only). | STRING |
model_task_id | The unique task ID for this model generation request. | MODEL_TASK_ID |
GLB | The generated 3D model in GLB format. | FILE3DGLB |
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub
Source fingerprint (SHA-256):
29bb87cdc5d3eef891a653c622e8876a37d6e6dc1a43e5c248b184060ead9029