ComfyUI Node
H3 • Text to Image
Describe one image and choose its size. Prepares exactly one H3 latent frame; no frame count or selection is needed.
H3 • Text to Image
- models
- image_job
◄prompt►
◄aspect_ratio1:1 square►
◄megapixels3.0►
◄width2048►
◄height1536►
CategoryMiniMax H3/Image Studio
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| models | H3_STUDIO_MODELS | Connect H3 Models directly, or the last optional H3 LoRA. | |
| prompt | STRING | Describe the final image. Your prompt is used directly without hidden preservation wording. | |
| aspect_ratio | COMBO | 1:1 square | Choose the canvas shape. Source image keeps Picture 1's ratio; custom pixels uses the advanced width/height controls. |
| megapixels | FLOAT | 3.00.1–32 | Target image size, rounded to a 32-pixel grid. Start at 3 MP; 1 MP is a preview, 4–8 MP costs more. Ignored in custom pixels mode. |
| widthopt | INT | 204864–16384 | Exact pixel width, only when aspect_ratio is custom pixels. |
| heightopt | INT | 153664–16384 | Exact pixel height, only when aspect_ratio is custom pixels. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| image_job | H3_STILL_JOB | Prepared still and cached prompt conditioning. Connect to Render Image. |