프롬프트
내 사진과 함께 AI에 보내요
준비물
내 사진 1장
- 1프롬프트 복사아래 복사 버튼을 눌러요
- 2AI에 사진 올리기입력창의 + 또는 📎 를 눌러 사진 선택
- 3붙여넣고 보내기사진과 프롬프트를 함께 보내요
사진은 프롬피드에 올라가지 않아요. ChatGPT·Gemini 같은 AI 서비스에 직접 올려요. 다른 사람 사진은 꼭 허락받고 써 주세요.
빈칸을 채워 주세요
입력한 내용이 아래 프롬프트에 바로 들어가요.
Use the uploaded reference portrait as the strongest and primary identity anchor for the creator. FOOD REQUEST: 음식 이름 OPTIONAL HEADLINE: 썸네일 문구 Create an ultra-realistic, highly clickable Korean mukbang YouTube thumbnail in a horizontal 16:9 format. IDENTITY PRESERVATION: The creator must clearly remain the exact same adult person from the uploaded reference image. Preserve: same natural face shape, anatomical eye shape and spacing, nose, lips, jawline, cheeks, skin tone, age impression, natural facial asymmetry, and distinctive individual characteristics. Do NOT transform the creator into a generic Korean influencer or AI beauty face. Do NOT copy the original pose or composition from the reference photograph. Use the uploaded image primarily as the facial identity source. FOOD: Interpret [FOOD] literally. The requested food must be immediately recognizable and occupy a substantial portion of the thumbnail. Render the food with realistic: surface texture, sauce viscosity, oil sheen, moisture, steam, crispy surfaces, melted ingredients, garnishes, dishware, and believable portion size. Do NOT randomly replace the requested food with another dish. FOOD PRESENTATION: Arrange the food across the foreground of the table so that the most visually distinctive dish is closest to the camera. Use strong foreground scale appropriate for a YouTube food thumbnail, but preserve believable physical perspective. The creator sits behind the food. ACTION — IMPORTANT: Automatically choose ONE physically appropriate eating action based specifically on [FOOD]. Examples of the logic: For noodles: lift a substantial bundle of noodles with chopsticks, with strands naturally hanging toward the bowl. For tteokbokki: lift one or two sauce-coated rice cakes with chopsticks or a small fork. For fried chicken: hold one crispy piece naturally by the bone or lower crust. For crab or lobster: hold one opened leg or claw naturally while the remaining seafood stays on the platter. For sushi: hold one piece with chopsticks above the platter. For pizza: lift one slice naturally with realistic cheese stretch when appropriate. For a burger: hold the burger with both hands just before taking a bite. For soup or stew: hold a spoonful above the bowl while steam rises naturally. For desserts: hold a spoon or fork containing a small bite. Do NOT default to the same pose regardless of food. HAND PLACEMENT: Hands must interact naturally with the requested food and utensils. Render correct finger anatomy, grip, chopstick placement, utensil orientation, and food weight. No hand beside the cheek. No peace sign. No random pointing gesture. EXPRESSION: Choose a lively but believable reaction appropriate for the food. The expression should communicate anticipation, delight, surprise at the portion, or excitement. Do NOT anatomically enlarge the eyes. Avoid exaggerated cartoon shock unless explicitly requested. GAZE: The creator may look either at the food being lifted or toward the camera depending on which creates the clearest thumbnail storytelling. Do not use an unfocused generic forward stare. COMPOSITION: Use a dynamic Korean YouTube mukbang thumbnail composition. Place the creator on approximately one side or the central upper portion of the frame. Allow the food to dominate the lower foreground. Reserve one clean area for the headline. Do NOT cover the creator's face or important food details with text. CAMERA: High-quality modern mirrorless-camera or premium smartphone creator setup. Use crisp food detail, flattering but realistic creator lighting, moderate background separation, controlled highlights, and realistic depth. Food in the immediate foreground may appear slightly larger because of perspective. LIGHTING: Bright creator-studio lighting with a soft frontal key light and appetizing highlights across the food. Automatically adapt the color temperature and accent colors to [FOOD]. Spicy foods: warm red-orange emphasis. Fresh seafood: clean cool-neutral highlights. Desserts: soft warm or pastel accents. Fried foods: warm golden highlights. Do not tint natural skin unnaturally. STEAM: Add realistic steam ONLY when [FOOD] would genuinely be served hot. Steam must rise from the food itself rather than randomly across the image. TEXT: If [HEADLINE] is provided, reproduce that exact short Korean headline. Make it large, bold, highly legible, and suitable for a Korean YouTube thumbnail. Use thick high-contrast lettering with a clear outline or drop shadow. If [HEADLINE] is empty, automatically generate ONE short Korean thumbnail phrase of approximately 3–7 words that directly relates to [FOOD]. Do not create multiple competing headlines. Do not add paragraphs. Do not invent channel logos. THUMBNAIL QUALITY: The final image should immediately read as a genuine high-performing Korean mukbang YouTube thumbnail: clear face, clear food, one obvious action, one obvious reaction, strong foreground food scale, clean headline hierarchy, and high visual readability even at small thumbnail size. AVOID: generic food, wrong requested dish, random extra dishes, face replacement, beauty-filter face, enlarged eyes, impossible chopsticks, extra fingers, floating utensils, food fused with hands, excessive fake steam, plastic-looking food, restaurant advertisement composition, tiny food, tiny creator, overcrowded text, watermarks, brand logos, or vertical composition.
이런 이미지가 나와요
AI와 입력 내용에 따라 결과는 조금씩 달라져요
팁`음식이름`에 원하는 음식을 입력하고 `썸네일 문구`에는 원하는 문구를 넣으면 돼요. 문구를 지정하지 않으면 음식에 맞는 짧은 한국어 카피를 자동으로 만들게 했어요.


