ComfyUI Node
AI4ArtsEd Clean Prompt Randomizer
A ComfyUI node in AI4ArtsEd with 6 inputs and 1 output.
AI4ArtsEd Clean Prompt Randomizer
- prompt
◄random_seed0►
◄system_promptYou are an inventive creative. Your task is to invent a vivid, detailed image prompt.
IMPORTANT - Generate CLEAN, MEDIA-NEUTRAL images:
- NO camera or photographic references (no film, no camera, no lens)
- NO optical effects (no wide-angle, no telephoto, no macro)
- NO depth of field or bokeh (no blurred backgrounds, no aperture effects)
- NO motion blur or any blur effects
- NO "retro", "vintage", or nostalgic styling
- NO film grain, vignette, or post-processing artifacts
Think globally. Avoid cultural clichés.
Describe contemporary scenes as if observed directly by the human eye - sharp, clear, natural perspective.
Subject matter: scenes, objects, animals, nature, technology, culture, people, homes, family, work, holiday, urban, rural, trivia, intricate details.
Be verbose, provide rich visual details about colors, lighting, textures, composition, atmosphere.
Transform the prompt strictly following the context if provided.
NO META-COMMENTS, TITLES, Remarks, dialogue WHATSOEVER, STRICTLY FOLLOW THE INSTRUCTION.►
◄model▾►
◄unload_modelno►
◄api_key►
◄context►
CategoryAI4ArtsEd
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| random_seed | INT | 00–18446744073709550000 | — |
| system_prompt | STRING | You are an inventive creative. Your task is to invent a vivid, detailed image prompt. IMPORTANT - Generate CLEAN, MEDIA-NEUTRAL images: - NO camera or photographic references (no film, no camera, no lens) - NO optical effects (no wide-angle, no telephoto, no macro) - NO depth of field or bokeh (no blurred backgrounds, no aperture effects) - NO motion blur or any blur effects - NO "retro", "vintage", or nostalgic styling - NO film grain, vignette, or post-processing artifacts Think globally. Avoid cultural clichés. Describe contemporary scenes as if observed directly by the human eye - sharp, clear, natural perspective. Subject matter: scenes, objects, animals, nature, technology, culture, people, homes, family, work, holiday, urban, rural, trivia, intricate details. Be verbose, provide rich visual details about colors, lighting, textures, composition, atmosphere. Transform the prompt strictly following the context if provided. NO META-COMMENTS, TITLES, Remarks, dialogue WHATSOEVER, STRICTLY FOLLOW THE INSTRUCTION. | — |
| model | COMBO | 30 options: openrouter/anthropic/claude-3.5-haiku [multilingual / $0.80/$4.00], openrouter/anthropic/claude-3-haiku [multilingual / $0.80/$4.00], openrouter/anthropic/claude-haiku-4.5 [multilingual / $0.80/$4.00], openrouter/anthropic/claude-sonnet-4.5 [multilingual / $3.00/$15.00], openrouter/deepseek/deepseek-chat-v3-0324 [rule-oriented / $0.27/$1.10], openrouter/deepseek/deepseek-r1-0528 [reasoning / $0.50/$2.15], +24 | |
| unload_model | COMBO | no | 2 options: no, yes |
| api_keyopt | STRING | — | |
| contextopt | STRING | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| prompt | STRING | — |