ComfyUI Node
Ideogram 4 Autoprompter
AI autoprompter for Ideogram 4's structured JSON caption format. Describe an idea (and/or drop a reference image) in the node UI and let a local vision LLM (Qwen3-VL, default) or Gemini build the full caption — background, placed elements with bboxes, descriptions, text, and color palettes. Everything remains editable on the canvas afterward. At run time the node only outputs the assembled caption JSON string (plus a preview).
Ideogram 4 Autoprompter
- prompt
- preview
◄width1024►
◄height1024►
◄high_level_description►
◄background►
◄style▾►
◄aesthetics►
◄lighting►
◄medium►
◄style_palette_data►
◄elements_data►
◄ai_state►
CategoryIdeogram/text
Inputs (11)
| Name | Type | Default | Description |
|---|---|---|---|
| width | INT | 102464–16384 | Canvas aspect width (also the pixel grid the bbox is measured in). |
| height | INT | 102464–16384 | Canvas aspect height (also the pixel grid the bbox is measured in). |
| high_level_description | STRING | Optional one-line overview of the whole image (blank = omitted). | |
| background | STRING | Required scene background description. | |
| style | COMBO | 3 options: [object Object], [object Object], [object Object] | |
| aesthetics | STRING | Style descriptor (blank = omitted). | |
| lighting | STRING | Style descriptor (blank = omitted). | |
| medium | STRING | Style descriptor (blank = omitted). | |
| style_palette_data | STRING | Serialized style color palette from the editor (managed by the node UI). | |
| elements_data | STRING | Serialized regions from the editor (managed by the node UI). | |
| ai_state | STRING | Serialized autoprompter UI state (managed by the node UI; never the API key). |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| prompt | STRING | — |
| preview | IMAGE | — |