Anime Chibi Stickers

character_1: name: 添付画像1 description: | 可愛く、元気な印象。 character_2: name: 添付画像2 description: | 表情豊かで、Character1の隣や肩の上などにいて、同調または補足的に動作や表情をする。 style: background: 白 layout: 3列×3行 × 4セット（全36コマ） outline: 柔らかい線画、デフォルメ調 text_style: 手書き風、感情に応じた色付き（太字） atmosphere: 明るくかわいいLINEスタンプ風 panels: - position: 1セット目 左上 text: おはよ〜 text_color: 黄色 girl_pose: 手を振る元気な挨拶、笑顔 penguin_pose: 羽を振って一緒に挨拶 - position: 1セット目 中央上 text: ありがとっ！ text_color: オレンジ girl_pose: 軽くお辞儀、目を細めた笑顔 penguin_pose: 両手を合わせてぺこり - position: 1セット目 右上 text: おつかれさま〜 text_color: 青 girl_pose: 敬礼ポーズ、微笑み penguin_pose: 肩の上で敬礼 - position: 1セット目 左中 text: OKだよ〜 text_color: 黄緑 girl_pose: 親指を立てて笑顔 penguin_pose: にこにこ＋キラキラの星 - position: 1セット目 中央中 text: ごめーん text_color: 青 girl_pose: 困り顔＋両手合わせて謝罪 penguin_pose: 頭を下げてしゅんとする - position: 1セット目 右中 text: いま行くー！ text_color: 赤 girl_pose: 走っている、元気な笑顔 penguin_pose: 一緒に跳ねながらついていく - position: 1セット目 左下 text: よろしくね text_color: ピンク girl_pose: 丁寧にお辞儀、微笑み penguin_pose: 横でぺこりとお辞儀 - position: 1セット目 中央下 text: おやすみ〜 text_color: 紫 girl_pose: 枕を抱えて眠そうな顔、あくび penguin_pose: ナイトキャップを被ってウトウト - position: 1セット目 右下 text: またね！ text_color: オレンジ girl_pose: 大きく手を振って見送る penguin_pose: 寂しそうに小さく羽を振る - position: 2セット目 左上 text: えっ！？まじで？ text_color: 水色 girl_pose: 驚き顔（目を丸く、口をあける） penguin_pose: 驚き表情＋汗マーク - position: 2セット目 中央上 text: ちょっと～！！ text_color: オレンジ＋青 girl_pose: 両手を広げて叫ぶ、焦り顔 penguin_pose: 焦ってタジタジ（汗マーク） - position: 2セット目 右上 text: やったー！ text_color: 赤 girl_pose: 両手を上げてジャンプ、満面の笑み penguin_pose: 万歳して一緒に喜ぶ - position: 2セット目 左中 text: むかっ text_color: 赤 girl_pose: 頬を膨らませて腕組み、怒り顔 penguin_pose: プイッとそっぽを向く - position: 2セット目 中央中 text: ぴえん text_color: 青 girl_pose: ウルウルした瞳で泣きそう penguin_pose: 大号泣して地面を叩く - position: 2セット目 右中 text: じーっ text_color: グレー girl_pose: 壁の影から覗き込む、真顔 penguin_pose: 肩の上で双眼鏡を覗く - position: 2セット目 左下 text: わくわく♪ text_color: ピンク girl_pose: 頬を染めて期待に満ちた顔 penguin_pose: 目をキラキラさせて待機 - position: 2セット目 中央下 text: なるほど！ text_color: 黄色 girl_pose: ポンと手を打つ、ひらめき顔 penguin_pose: ウンウンと深く頷く - position: 2セット目 右下 text: それな text_color: 緑 girl_pose: ビシッと指を差す、ドヤ顔 penguin_pose: 横で全力の頷き - position: 3セット目 左上 text: がんばれー！ text_color: オレンジ girl_pose: 拳を握ってエールを送る penguin_pose: 旗を一生懸命振る - position: 3セット目 中央上 text: お願いっ text_color: ピンク girl_pose: 両手を合わせて上目遣い penguin_pose: 一緒に拝むポーズ - position: 3セット目 右上 text: 了解です！ text_color: 緑 girl_pose: ビシッと敬礼、キリッとした笑顔 penguin_pose: 女の子の頭の上で敬礼 - position: 3セット目 左中 text: 無理〜！！ text_color: 青 girl_pose: 両手でバツ印、白目をむく penguin_pose: 泡を吹いて倒れている - position: 3セット目 中央中 text: 大丈夫？ text_color: 黄緑 girl_pose: 覗き込むように心配そうな顔 penguin_pose: 背中をポンポン叩く - position: 3セット目 右中 text: ぱちぱちぱち text_color: オレンジ girl_pose: 笑顔で拍手、キラキラマーク penguin_pose: 羽を叩いて一緒に拍手 - position: 3セット目 左下 text: おめでとう！ text_color: 赤 girl_pose: クラッカーを鳴らす、大喜び penguin_pose: 紙吹雪をまいて祝福 - position: 3セット目 中央下 text: 楽しみ〜！ text_color: ピンク girl_pose: スキップしている、楽しそうな顔 penguin_pose: 浮き輪を持って準備万端 - position: 3セット目 右下 text: さすが！ text_color: 黄色 girl_pose: 両手で親指を立てる penguin_pose: 「NO.1」の看板を掲げる - position: 4セット目 左上 text: はらぺこ… text_color: 茶色 girl_pose: お腹を押さえて力尽きそう penguin_pose: 空の皿とスプーンを持って待機 - position: 4セット目 中央上 text: もぐもぐ text_color: オレンジ girl_pose: 頬張って幸せそうな顔 penguin_pose: 一緒に魚（好物）を食べている - position: 4セット目 右上 text: 休憩中〜 text_color: 緑 girl_pose: お茶を飲んでホッとする penguin_pose: 湯呑みを抱えてのんびり - position: 4セット目 左中 text: チラッ text_color: グレー girl_pose: 画面の端からひょっこり出す penguin_pose: 女の子の下からひょっこり - position: 4セット目 中央中 text: 逃げろー！ text_color: 青 girl_pose: 砂煙を上げて全力疾走 penguin_pose: 必死に後ろからついていく - position: 4セット目 右中 text: 準備中… text_color: 黄色 girl_pose: 鏡の前で身支度、真剣な顔 penguin_pose: ブラシを持って手伝うふり - position: 4セット目 左下 text: 到着っ！ text_color: 赤 girl_pose: 目的地でポーズ、達成感のある顔 penguin_pose: 地図を広げてピース - position: 4セット目 中央下 text: ぐったり text_color: 青 girl_pose: 床にのの字を書いて落ち込む penguin_pose: 氷の嚢を頭に乗せてダウン - position: 4セット目 右下 text: ぺこり text_color: 黒 girl_pose: 丁寧にお辞儀、感謝の表情 penguin_pose: 帽子を脱いで一緒にお辞儀

[SOURCE MODE = ATTACHED IMAGE] Create a image 16:9 experimental graphic composition using a recognition-preserving annotated pixel-slice deconstruction system. SOURCE PRIORITY The original source is the dominant master layer. The analytical grid, annotations and pixel-stretch effects are secondary overlays applied around and selectively across the source. Never reconstruct the subject from separate grid fragments. Never turn the source into a collage of disconnected crops. The complete original subject must remain immediately recognizable at first glance before the viewer notices the analytical distortion system. If [SOURCE MODE = ATTACHED IMAGE], use the uploaded image as the sole visual source. Preserve its actual subject, silhouette, proportions, pose, orientation, perspective, internal geometry, colors, material appearance and major recognizable features. If [SOURCE MODE = LOGO], automatically identify the authentic official standalone symbol of [LOGO NAME]. Preserve its canonical silhouette, proportions, curves, stroke relationships, counterforms, internal negative spaces, intersections and component spacing. Do not redesign, simplify or reinterpret the mark. RECOGNIZABILITY LOCK Preserve approximately 60–75% of the visible source as continuous, undistorted imagery. At least one large uninterrupted region must show most of the source in its original spatial arrangement. The viewer must be able to mentally reconstruct the entire source without effort. Protect all identity-critical geometry from destructive distortion: — outer silhouette; — primary junctions; — distinctive corners and curves; — central intersections; — negative spaces; — major directional axes; — defining internal relationships between parts. Do not place heavy pixel stretching across these critical recognition anchors. If distortion would weaken recognition, reduce the distortion rather than modifying the source geometry. The source should remain spatially coherent underneath the graphic system. CORE VISUAL SYSTEM Build an irregular architectural grid of approximately 18–30 rectangular analytical cells around and partially through the source. The grid should feel like an experimental scientific image-analysis interface, dataset annotation sheet and corrupted editorial layout merged into one system. The source remains the hero. The grid must behave like a layer placed over and around an already complete image, not like a structure used to rebuild it. Allow the grid to extend into surrounding negative space, intersect the source at selected locations and isolate small analytical samples from it. Use a few large anchor cells, several medium cells and narrow micro-cells with asymmetric stepped relationships. Most cells should share boundaries or connect structurally. Do not use an even checkerboard. Do not cover the entire source with grid divisions. Do not force every visible part of the source into a separate rectangle. SOURCE CONTINUITY Where grid lines cross the undistorted source, the image underneath must continue seamlessly across adjacent cells. A grid boundary should not automatically create a new crop. The same original object must remain positionally aligned from one cell to the next. Do not shift, rescale, rotate or independently reposition adjacent source fragments. Do not duplicate distinctive parts of the source unless intentionally used as a small analytical sample. The grid is graphic segmentation, not physical fragmentation. PIXEL-SLICE DECONSTRUCTION Apply pixel stretching only to selected secondary cells. Use approximately 20–35% of the grid cells for strong distortion. Sample very narrow strips directly from nearby regions of the source and stretch them across the selected cell. Create: — vertical source slices extended into vertical spectral bands; — horizontal source slices extended into horizontal bands; — compressed local color profiles; — elongated texture streaks; — repeated edge-color information; — smooth raster smear fields derived directly from the source. The distorted sample must remain visually related to the exact neighboring area it was extracted from. Whenever possible, place stretched cells directly beside the intact region they analyze, creating a clear visual relationship between source and sample. Do not replace a large recognizable portion of the source with distortion. Do not stretch the entire object. Do not let distorted cells become more visually dominant than the original source. IMAGE CELL HIERARCHY Use three clear cell types: 1. PRIMARY SOURCE CELLS Large cells containing continuous undistorted portions of the original source. These carry most of the recognizability. 2. ANALYTICAL SAMPLE CELLS Smaller cells containing pixel-stretched or isolated source samples. 3. EMPTY INFORMATION CELLS Background-only cells containing a label, index number or nothing at all. Primary source cells should dominate the composition by area. Analytical sample cells should behave like extracted measurements surrounding the source. Empty cells create breathing room and prevent the composition from becoming a dense mosaic. ANNOTATION SYSTEM Give selected cells tiny technical annotations positioned near the upper-left corner. Automatically derive concise semantic labels from the actual source: SUBJECT SILHOUETTE CURVATURE EDGE JUNCTION REFLECTION CONTOUR NEGATIVE SPACE MATERIAL GRADIENT STRUCTURE COLOR FIELD VERTICAL SAMPLE HORIZONTAL SAMPLE INTERSECTION PROFILE Use uppercase condensed grotesk or monospaced technical typography. Typography should be black or near-black on light backgrounds and adapt subtly for dark backgrounds. Keep all labels very small. Pair labels with sequential index numbers such as 1, 2, 3, 4… Occasionally use a tiny arrow, dash or technical symbol. Never allow text to become a headline or compete visually with the source. GRID GRAPHICS Use extremely thin neutral-gray or near-black hairline rules. Keep borders precise, flat and diagrammatic. Grid lines may cross the source, but crossing a grid line must not automatically distort or relocate the image. Allow some cells to extend completely into empty background space. The overall footprint should be asymmetric and architectural, with occasional stepped protrusions and isolated narrow analytical columns. BACKGROUND + NEGATIVE SPACE Use [BACKGROUND COLOR] as the surrounding field and empty-cell color. Keep approximately 40–60% of the total canvas visually quiet. Large areas of untouched background are essential. Do not surround the entire source with equally dense cells. Concentrate more analytical activity in two or three localized zones and allow other areas to remain calm. COMPOSITION Square 1:1 canvas. Preserve the original source orientation and major compositional balance. Keep the subject large enough to be recognized instantly. Do not aggressively zoom into isolated details at the expense of the whole object. The intact source should form one dominant visual mass. The analytical grid should form a secondary asymmetric constellation around and partially across it. The final composition should read in this order: 1. recognizable original subject; 2. analytical rectangular system; 3. pixel-stretch details; 4. micro-annotations. Never reverse this hierarchy. RENDER CHARACTER Extremely crisp 2D graphic design. Preserve the original photographic, rendered or vector-like character of the source inside undistorted regions. Combine it with precise raster stretching, exact rectangular clipping, thin diagrammatic rules and restrained technical typography. No added physical depth or material treatment beyond what already exists in the source. NEGATIVE CONSTRAINTS Do not rebuild the subject from independent rectangular fragments. Do not turn the source into a contact sheet. Do not create a mosaic. Do not destroy the overall silhouette. Do not split every part of the object into separate cells. Do not independently reposition adjacent source fragments. Do not duplicate major recognizable features. Do not crop away identity-critical regions. Do not let distorted samples occupy more visual area than the intact source. Do not create generic cyberpunk glitch art. No RGB channel separation. No VHS noise. No scanlines. No datamoshing. No JPEG artifacts. No random pixel blocks. No unrelated gradients. No arbitrary neon colors. No regular checkerboard. No uniform modular grid. No oversized typography. No decorative UI icons. No fake charts. No complete source abstraction. FINAL PRIORITY If any instruction conflicts with source recognizability, preserve recognizability first. The finished image must look like a complete original image being analytically sampled, segmented and partially stretched by an external graphic system — never like the original image was destroyed and reconstructed from glitch fragments.

Create a highly realistic studio photograph of a single natural stone monolith with the exact logo from IMAGE 1 shallowly carved into its front surface. IMAGE 1 = exact logo source. Preserve the overall logo shape, proportions and spacing exactly. IMAGE 2 = inspiration and visual reference. Match the smooth dense stone material, primitive carved feel, subtle dirt buildup and realism, but do not copy the exact stone or symbol. Stone color: [INSERT COLOR] The stone should look dense, smooth, compact and naturally weathered, with a relatively flat and smooth front face. Avoid porous, sandy, chalky, hammered, pecked or chisel-struck textures. The front face should feel slightly worn and subtly polished, while the outer sides and edges can remain rougher and more irregular. Carve the logo as a shallow hand-carved engraving into the smooth front face. Preserve the exact overall logo silhouette, but make the engraved edges slightly rounded, softly imperfect, subtly chipped and naturally worn at the micro level. The edges must not look sharp, laser-cut or mechanically precise. The recessed logo area should stand out only moderately from the rest of the stone. If the stone is dark, make the recessed area subtly lighter; if the stone is light, make it subtly darker. Add restrained dust, smudges, slight grime and small lighter worn spots near parts of the inner edges so the logo is more visible, but not artificially contrasted. Do not make the entire stone face look carved with repeated chisel impacts. Only the logo edges should show handcrafted wear. Pure black background. Premium directional studio lighting. Front-facing composition with slight perspective. Photorealistic, high-end artifact photography, 4K, Ultra Quality.

Create an ultra-photorealistic vertical selfie in 9:16 format. Use the TWO uploaded reference images as exact character references: one woman and one man. Preserve their identities with 100% accuracy. Do not change facial features, face shape, proportions, age, eye shape, nose, lips, eyebrows, skin tone, hairstyle, or any other defining characteristics. Their faces must remain completely recognizable and natural, without beautification or modification. Scene: The woman and man are sitting together on the edge of the highest metal platform of the Empire State Building in New York, at an extreme height above Manhattan. They are sitting close together on the exposed metallic structure, with the Manhattan skyline clearly visible far below. Both are looking directly into the camera with relaxed, confident expressions and subtle natural smiles. Pose: relaxed posture, natural weight distribution, asymmetrical pose. The man is sitting slightly behind the woman while the woman is closer to the camera. Their bodies should look naturally positioned. Hands and fingers must be anatomically correct. Camera: realistic smartphone selfie perspective, approximately 24mm wide-angle lens, slight perspective distortion similar to a real front-facing camera. Camera positioned slightly above eye level. Faces should be large and clearly visible in the frame. Composition: the two people and their faces are the main focus, but the extreme height and Manhattan skyline must remain clearly visible in the background. No artificial blur hiding the environment. The background should have realistic depth and detail. Environment: exposed steel platform and metallic construction of the Empire State Building, narrow areas of steel grating, beams, bolts, nuts, joints, metal framework, and a technical control box with several round buttons nearby. Everything should look physically realistic and structurally believable. Lighting: natural daylight, realistic skin texture, subtle shadows, cinematic but completely photorealistic. No excessive beauty filter, plastic skin, artificial glow, or exaggerated facial features. The final image must look like a real spontaneous smartphone selfie taken at an extreme height above New York City, with highly realistic people, realistic skin, accurate anatomy, detailed metal structures, and a convincing sense of altitude.

Create a vertical 3:4 commercial campaign poster for a fictional [Product Category] brand. Brand: [Brand Name] Product: [Product Name] Main word: [Main Word] Tagline: [Tagline] Color palette: [Color Palette] Model: [Model] Key features: [Feature 1], [Feature 2], [Feature 3], [Feature 4] Use oversized bold typography with a strong one-point or diagonal perspective. The main word must feel integrated into the scene as a giant wall, corridor, street, stage, or spatial structure, with clear depth and readable letterforms. Place the character near the center of the composition. The character must hold the product and extend that hand toward the camera, creating an exaggerated foreground and a strong 3D wide-angle effect. The product should be the closest, largest, and sharpest element, followed by the character in the midground and the giant typography in the background. Keep the product shape accurate, realistic, and clearly branded. Add appropriate details such as condensation, ice, bubbles, fruit, sauce, steam, or packaging elements, but keep them controlled and secondary. Include a complete campaign layout with brand name, feature icons, vertical side text, campaign number, release date, product details, barcode, and micro typography. Avoid a thick footer bar that covers the character; distribute the information across the lower corners and side margins. High-resolution commercial photography, realistic skin and materials, strong perspective, dynamic pose, polished brand design, high visual impact.

Use my uploaded reference image as the primary facial identity reference and preserve my exact facial identity with maximum accuracy. Keep my face shape, jawline, cheekbones, eyebrows, eyes, nose, lips, ears, skin tone, hairstyle, hair texture, beard/stubble pattern, and all distinctive facial features consistent and recognizable. Do not change, beautify, stylize, or replace my identity. Create a highly photorealistic cinematic portrait of me standing on an old traditional balcony, wearing a loose, oversized black T-shirt with relaxed sleeves and natural fabric folds. The T-shirt should have a premium heavyweight cotton texture and drape naturally over my body. I am standing confidently behind an aged wrought-iron balcony railing with decorative curved patterns. My body is slightly angled toward the camera while my head is tilted subtly upward and toward the side, creating a calm, mysterious and introspective expression. One hand rests naturally near my waist while the other remains relaxed. Keep my appearance masculine and natural, with realistic skin texture, visible pores, subtle facial imperfections, detailed hair strands and realistic beard texture. Add a simple silver chain necklace, subtle silver rings, bracelets, and a small metallic hoop earring. A cigarette rests naturally between my lips, producing soft wisps of realistic smoke drifting upward into the darker background. The background features an an old Indian-style residential building, with a large dark arched doorway, aged plaster walls, weathered architectural details and deep shadows. The balcony railing occupies the lower foreground and creates strong depth between the camera and subject. Use soft warm natural daylight from the side, illuminating my face and shoulders while maintaining deep cinematic shadows around the doorway. Add subtle atmospheric haze and realistic volumetric smoke. Camera: vertical 4:5 portrait, eye/chest-level perspective with a slight low-angle feel, 85mm portrait lens, f/2–f/2.8, shallow depth of field, sharp focus on my face and upper body, softly blurred background. Visual style: ultra-photorealistic, cinematic editorial photography, natural skin tones, realistic fabric and metal textures, detailed beard and hair, physically accurate lighting, subtle film grain, rich tonal range, muted warm colors, dramatic but natural shadows, premium DSLR/medium-format photography, 8K detail. Negative prompt: different person, altered identity, face distortion, excessive beautification, plastic skin, CGI, cartoon, artificial muscles, unrealistic anatomy, deformed hands, extra fingers, warped jewelry, unnatural smoke, oversaturated colors, excessive HDR, blurry face, low detail, watermark, text, logo.

{ "prompt_template": { "subject": "Person Name/Uploaded Image", "style": "hyperrealistic studio portrait photograph, cinematic dual-tone lighting", "composition": { "framing": "chest-up close crop portrait", "position": "perfectly centered, symmetrical", "orientation": "vertical, portrait aspect ratio 4:5", "gaze": "direct eye contact with camera, head facing forward, chin level" }, "expression": { "mood": "calm, neutral, slightly serious", "mouth": "closed, relaxed lips", "eyes": "sharp, alert, direct" }, "wardrobe": { "clothing": "plain oversized white crewneck t-shirt or sweatshirt", "details": "no logos, no patterns, relaxed loose fit, shoulders fully visible", "color": "pure white, slightly glowing under fill light" }, "lighting": { "setup": "dual-tone split lighting", "key_light": "warm amber/orange rim light from camera-left, raking across one side of face, neck and shoulder", "fill_light": "cool blue ambient light filling the rest of the face and body", "background_light": "soft glow gradient matching blue ambient tone", "contrast": "high contrast between warm highlight and cool shadow zones", "quality": "soft studio diffusion with a crisp rim highlight, no harsh shadows on facial features" }, "background": { "type": "seamless studio backdrop", "color_gradient": "deep navy blue to teal, darker at edges", "vignette": "subtle darkening toward corners", "texture": "smooth, no props, no patterns" }, "camera_specs": { "lens": "85mm equivalent portrait lens", "aperture": "shallow depth of field, f/2.8 look", "focus": "tack sharp on eyes and facial features", "sensor_feel": "medium format digital, ultra high resolution, 8k detail" }, "skin_and_hair_detail": { "skin": "natural realistic texture, visible pores, subtle sheen from lighting, no over-smoothing or plastic look", "hair": "natural texture and volume matching subject's real hairstyle, individual strands visible under rim light" }, "color_grade": "cinematic teal and orange color grade, warm highlights vs cool shadows, filmic contrast", "post_processing": "professional retouch level, subtle skin evenness, retain natural imperfections, no beautify filter, no plastic AI look", "negative_prompt": "no text, no watermark, no extra objects, no inset image, no reference photo overlay, no multiple people, no distorted face, no blurry eyes, no cartoon or illustration style, no harsh unnatural shadows", "output_settings": { "resolution": "high resolution, ultra detailed 8k", "aspect_ratio": "4:5 portrait", "realism_level": "photorealistic, indistinguishable from a real studio photograph" } } }

Use my uploaded face image as the primary facial identity reference. Preserve my exact facial identity with maximum accuracy. My face must remain instantly recognizable, including my face shape, jawline, hairstyle, hair texture, eyebrows, eye shape, nose, lips, ears, skin tone, beard, facial proportions, and every unique facial feature. Do not beautify, stylize, age, de-age, or replace my identity. A cinematic red carpet arrival at a prestigious movie premiere during the evening. I am confidently walking down a luxurious red carpet, smiling warmly while raising one hand to wave at a massive crowd of enthusiastic fans standing behind black barricades on both sides, many holding up smartphones to capture the moment. Several professional security personnel in elegant black suits follow a few steps behind, maintaining a respectful distance. I wear an oversized premium burgundy/maroon crew-neck T-shirt with a subtle minimalist chest logo, wide-leg black tailored trousers, sleek black slip-on sneakers with white soles, a luxury wristwatch, and one hand casually tucked into my pocket. My hairstyle is voluminous and naturally textured, with a neatly groomed beard. The atmosphere feels like a blockbuster film premiere with bright event lighting, warm cinematic highlights, realistic shadows, glowing bokeh from distant city lights, photographers, excited fans, and a glamorous high-profile celebrity ambiance. Camera: Full-body portrait, eye-level perspective, centered composition, 85mm lens, f/2.0, shallow depth of field, natural perspective, slight walking motion captured sharply. Lighting: Warm red-carpet spotlights mixed with soft ambient city lighting, realistic skin tones, subtle rim lighting around the subject, high dynamic range. Style: Ultra-realistic, Hollywood celebrity photography, luxury fashion editorial, premium event photography, cinematic color grading, highly detailed fabric textures, natural facial expressions, realistic crowd depth, 8K, HDR, photorealistic, magazine-quality, crisp focus on the subject with softly blurred background.