#2/2 How to Prompt the Anima AI Model: A Complete Guide
6
52
The following prompt was tested on a merged version of Anima and produced strong results on both short and long prompts. It demonstrates the full range of techniques: stacked quality tags, weighted artist references, layered descriptors, and a rich atmospheric finish.
masterpiece, best quality, amazing quality, newest, very aesthetic, absurdres, score 9, score 8, safe,
ultra detailed, highly detailed, refined details,
fantasy illustration, painterly, no lineart, lineless, soft painting, color focused, soft edges,
delicate rendering, soft bloom, depth of field, dreamy_bokeh, cinematic lighting, atmospheric perspective,
sharp focus, eye focus,
(@zocehuy:1.2), (@zomzomzomsauce:1.0),
1girl, solo, close-up portrait, extremecloseup, face focus, eye focus, lookingatviewer,
blue eyes, detailed eyes, reflective eyes, glossy eyes, long eyelashes, delicate eyelashes,
pale skin, translucent skin, glossy skin, soft blush, delicate features, elegant expression,
black hair, wispy bangs,
a large blue peony flower covering half of her face, flower touching her cheek,
petals blending into the composition, surrounded by soft blue flowers fading into the foreground and background,
monochromatic blue color palette, icy blue lighting, floral framing, delicate flower petals,
elegant earrings, subtle makeup, soft highlights around the eyes,
dreamlike floral atmosphere, ethereal beauty, fine art portrait photography aesthetic,
luxury fashion editorial feeling, soft color transitions, painterly textures,
magical realism, emotional composition, serene atmosphere, luminous skin,
subtle reflections, gentle shadows,
blue theme, monochrome blue, peony, flowers, floral background, flower frame,
blurry foreground, blurry background, shallow depth of field, atmospheric depth,
dreamy, ethereal, elegant, fantasyart, cinematic depth, soft_lighting
What makes this prompt work:
- Quality block first stacking masterpiece, amazing quality, score9, score8 alongside newest and absurdres pushes the model toward its highest fidelity output.
- Render style tags painterly, no lineart, lineless, soft painting explicitly tell the model to avoid harsh outlines and aim for an illustrated, brushed look.
- Weighted artists (@zocehuy:1.2) takes the dominant role in style while (@zomzomzomsauce:1.0) adds a secondary layer.
- Layered subject description eyes, skin, hair, and expression are each described separately before moving to the compositional element (the peony flower).
- Natural language for mood phrases like "a large blue peony flower covering half of her face" add compositional intent that pure tags can't convey.
- Closing tag cluster the final block bluetheme, dreamy, cinematicdepth) reinforces the overall aesthetic at the end, which helps the model stay consistent in later denoising steps.
Full Basic Tag Example
Here is a complete example of a simpler tag-style prompt: year 2025, newest, normal quality, score_5, highres, safe, 1girl, oomuro sakurako, yuru yuri, @ nnn yryr, smile, brown hair, hat, solo, fur-trimmed gloves, open mouth, long hair, gift box, fang, skirt, red gloves, blunt bangs, gloves, one eye closed, shirt, brown eyes, santa costume, red hat, skin fang, twitter username, white background, holding bag, fur trim, simple background, brown skirt, bag, gift bag, looking at viewer, santa hat, ;d, red shirt, box, gift, fur-trimmed headwear, holding, red capelet, holding box, capelet
Generation Settings
A variety of samplers work with Anima. Recommended ones: ersde for neutral style, flat colors, and sharp lines a good default. eulera produces softer, thinner lines and can tend toward a 2.5D look. dpmpp2msdegpu is similar to ersde but can produce more variety, though it can get unpredictable depending on the prompt. Recommended: 30–50 steps, CFG 4–5, resolution around 1024×1024.
