Nano Banana의 대화형 이미지 생성 방식에 맞춰 주체·행동·장소·구도와 선택한 스타일을 자연어로 연결합니다.
Generate an image of {{subject}} in {{scene}}. Establish a clear visual story through the subject's appearance, expression or action, its relationship to the location, and the important foreground, middle-ground, and background elements.
Use this selected art direction consistently: {{style_preset}}
Compose and photograph or render the scene as follows: {{camera}}. Make the main subject immediately readable, maintain natural spatial relationships, and keep lighting, perspective, reflections, and shadows internally consistent.
Follow these requirements exactly: {{constraints}}. If reference images are provided, preserve the referenced identity, proportions, product geometry, wardrobe, palette, and all details that are not explicitly changed. Do not add unrequested text, logos, duplicate subjects, or decorative objects. Produce one cohesive finished image rather than a collage of unrelated style effects.identity drift, inconsistent face, altered age, malformed anatomy, extra fingers, fused hands, duplicate subject, warped product geometry, invented controls, illegible text, logo, watermark, accidental crop, cluttered silhouette, flat lighting, low-detail textures사진, 회화, 애니메이션, 그래픽, 3D, 공예, 포스터, 건축과 콘셉트 아트 계열을 제공합니다. 선택값은 Nano Banana가 이해하기 쉬운 자연어 스타일 문장으로 삽입됩니다.
참조 인물·제품의 일관성, 장면의 공간 관계, 텍스트 철자, 원근, 반사와 그림자를 확인하세요.