Nodes/Multimodal — LLM (Local-first: GGUF + Transformers)/Bonus: LLM text → image intent (KSampler prompts)
ComfyUI Node
Bonus: LLM text → image intent (KSampler prompts)
**Bonus:** Detects phrases like “generate an image of …”, “draw a picture …”. Outputs **positive_prompt** / **negative_prompt** for your CLIP Encode → **KSampler** chain. **run_image_gen** is True when intent matches; use with any Switch/reroute node. English keyword heuristics only; tune **positive_suffix** for style.
Bonus: LLM text → image intent (KSampler prompts)
- run_image_gen
- positive_prompt
- negative_prompt
- parse_note
- run_image_gen_int
◄text►
◄negative_promptworst quality, low quality, blurry, watermark, text, logo►
◄positive_suffixhigh quality, detailed, coherent►
◄strip_thinking_tagstrue►
Categoryllm/multimodal
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| text | STRING | — | |
| negative_prompt | STRING | worst quality, low quality, blurry, watermark, text, logo | — |
| positive_suffix | STRING | high quality, detailed, coherent | — |
| strip_thinking_tags | BOOLEAN | true | — |
Outputs (5)
| Name | Type | Description |
|---|---|---|
| run_image_gen | BOOLEAN | — |
| positive_prompt | STRING | — |
| negative_prompt | STRING | — |
| parse_note | STRING | — |
| run_image_gen_int | INT | — |