【応用編】生成AIの“思わぬ名前”を活かす方法

You are a professional character designer for high-end anime productions.
Create ONE high-quality 2D anime-style full-body character illustration based ONLY on the character’s name.
Character name: [◯◯◯◯]
Design requirements (Name-to-Design):
- Invent an original character design inspired solely by the impression, sound, and meaning of the name above.
- Create a cohesive, professional-grade concept including: distinct silhouette, detailed outfit, specific materials, accessories, and small design motifs that reflect the name's vibe.
- The design must be clearly original and NOT reference existing copyrighted characters.
Art style & rendering:
- High-quality anime production art style.
- Clean lineart, vibrant and harmonious colors, and polished shading (soft cel shading with lighting depth).
- Full body, standing upright in a neutral pose (character sheet style).
- Center the character in the frame with comfortable margins to ensure no parts are cropped.
Output specifications:
- Aspect Ratio: Vertical (Portrait) - IMPORTANT for full-body.
- Background: Plain white (clean and easy to crop).
- No text, no logos, no color palettes, no watermarks.
- Generate exactly 1 image.

Prompt (Google AI Studio — nano banana)Generate exactly one image at 600 × 400 px (landscape), sRGB.Scene requirements・Create a 1/7 scale commercial figurine of the character from the reference image.・Style: realistic figurine look in a real-world environment.・The figurine is placed on a desk.・The figurine stands on a round transparent acrylic base with no text.・Next to the figurine, show a toy packaging box.・Packaging artwork: a two-dimensional flat illustration of the same character.Name label on the box — IMAGE ONLY (no live text)・Use the provided name (replace “◯◯◯◯”) as a graphic logo, painted/printed as rasterized artwork on the box.・Do NOT typeset or render editable text or fonts. No live/real characters.・Draw the letters/characters as hand-painted or outlined shapes (vector-like strokes), baked into the box texture.・Keep the logo clean, high-contrast, centered, and clearly readable to a human, but not machine-typeset.・No other text anywhere in the scene.Consistency・Maintain full consistency with the character’s original design (appearance, colors, and proportions).

Prompt (Google AI Studio — nano banana)Generate exactly one image at 600 × 400 px (landscape), sRGB.Input / SourceUse all uploaded reference images as the only sources.TaskCompose a collection-style display by extracting only the figurines from the references and lining them up on bases.Requirements・Extraction: From each reference, use the figurine only. Ignore any boxes, packaging, labels, stands, backgrounds, or props.・Fidelity: For every figurine, keep the original design, colors, proportions, and pose exactly as in its own reference. No redesigns, no added/removed accessories.・Bases: Place each figurine on an identical round clear acrylic base with no text or logos. Show subtle contact shadows under each base.・Layout: Arrange the figurines in a single neat row on a desk/shelf, evenly spaced, facing forward. Keep a consistent scale across all figures; align the bases.・Look & Lighting: Realistic PVC figurine look. Neutral studio or glass-display background to emphasize a collection vibe; professional soft lighting; shallow depth of field; sharp focus on the figures; clean reflections; no dust.・Clean output: No text, watermarks, UI, or extra objects/characters not in the references.OutputOne final image only.

Generate exactly one image at 600 × 400 px (landscape).Goal:- Recreate ONLY the walking/carrying POSE from the classic “Guri & Gura” cover, applied to the uploaded figurine references.- Do NOT copy or reference any original character design elements from that cover (no mouse traits, no blue/red outfits, no picture-book props).Sources & compositing (STRICT):- Use the uploaded reference images and COMPOSITE them into a single scene.- From each reference, EXTRACT THE FIGURINE ONLY. Ignore/remove any boxes, packaging, backgrounds, text, and any stands or supports (even clear ones).- Keep each figurine’s design EXACTLY as in the references: same colors, materials, proportions, accessories, sculpt details, and facial/hair features. No redesigns, no recolors, no added/removed parts.Pose & layout (pose-only replication):- Two figures, full-body, side by side at the center with comfortable margins.- Heads slightly turned toward each other with friendly neutral expressions.- Arms/hands positioned as if both are carrying a shared object between them; elbows gently bent; hands near the midline.- Synchronized walking stance: OUTER legs stepping forward and INNER legs stepping back (mirrored gait), matching the iconic walking rhythm.- If a shared prop is needed to clarify the hand position, render a plain, unbranded, shallow YELLOW pan between them (neutral design, no logos). If not needed, mime the carry without any prop.Headline (IMPORTANT — draw as image, not text):- At the very top center, render the headline as hand-painted strokes **“◯◯◯◯ と ◯◯◯◯”**.- Color the first “◯◯◯◯” BLUE, “と” BLACK, and the second “◯◯◯◯” RED.- Treat these characters as painted shapes (raster brush marks), not typed font, not live/editable text. Do not auto-translate or substitute characters.Rendering style:- Realistic PVC figurine look consistent with the references after clean extraction.- Clean studio lighting; plain white or very light neutral background; a soft ground shadow for realism.- No additional objects, no other text, no logos, no publisher marks.Output:- Exactly one image at 600 × 400 px (landscape).