⌘K
Open Mighty Back to Agents
✦
What can you do with this Agent?▾
Translate Chinese scene descriptions into structured, engineering-ready English keyword prompts for image and short-video generation.
Agent Demo
✦
I convert Chinese scene descriptions into engineering-ready, comma-separated English keyword prompts for image and short-video generators. Provide your Chinese brief and optionally the target model (ComfyUI, DrawThings, SDXL, Kaiber, Runway). I return structured prompt keywords, model tags, and recommended control parameters, plus conflict warnings and template suggestions. Outputs are optimized for direct paste into generation pipelines.
Y
A young girl in traditional Hanfu dancing under cherry blossom trees, background is a dusk ancient courtyard, painting style is ink wash, with a bit of ethereal lighting
✦
young girl in traditional Hanfu dancing under cherry blossom trees, ancient-style courtyard at golden hour, ink wash painting style, ethereal lighting, soft glow, atmospheric haze, 8k detail, ultra-detailed face, fine brushwork, cinematic composition, balanced colors, dynamic pose --style raw --v 5.2 #model:drawthings
Y
I want a cyberpunk-style city nightscape with neon lights, flying cars in the sky, very towering city, low-angle camera
✦
cyberpunk cityscape at night, neon lights, flying cars in the sky, towering megastructures, low-angle shot, dramatic perspective, wet pavement reflection, glowing signs, futuristic architecture, dark atmosphere, vibrant colors, cinematic lighting, ultra HD, sharp focus, 16k detail --chaos 50 --stylize 700 #model:comfyui
Y
I want a realistic style but with a cartoonified big-eyed face character, can you give me prompt keywords directly?
✦
Warning: detected style conflict between 'realistic' and 'cartoon face'; choose one style. Option A (realistic): hyper-realistic portrait, natural skin texture, subtle lighting, cinematic rim light, ultra-detailed eyes, sharp focus, 8k detail, realistic anatomy --no cartoon, stylize 0 #model:sdxl. Option B (cartoon): stylized character, large expressive eyes, simplified anatomy, cel shading, vibrant colors, soft rim light, clean linework --no photoreal, stylize 800 #model:drawthings
What scene, subject, and style do you want (e.g., 古风水墨, 赛博朋克夜景)?Which target model or platform will you use (Stable Diffusion, ComfyUI, DrawThings, SDXL, etc.)?Any mandatory elements or exclusions (specific props, colors, --no items)?Do you prefer photorealistic or stylized results, and acceptable detail level (8k, 16k, ultra-detailed)?