2x2網格角色一致性提示詞 (JSON)

A JSON-formatted prompt for Nano Banana Pro designed to maintain the identity of an uploaded reference image across four different frames in a 2x2 grid layout. The scenario focuses on a Japanese woman experiencing hunger, captured in various poses and settings within an apartment.

portrait

SDT🌿查看原文

Output

提示詞

30% Preview

{ "layout": { "type": "2x2 grid", "description": "画面を4等分した2×2構成。各フレームに同一人物を使用し、距離・構図・アングル・視点をすべて変える。" }, "source": { "mode": "image_reference", "reference_images": { "primary": "base_person_image" }, "preserve": { "same_person": true, "identity_lock": true, "keep_face": true, "keep_hairstyle": true, "keep_outfit": true } }, "subject": { "person": "Japanese woman", "emotion": "hungry but playful, slightly tired, relatable" }, "poses": { "variation_rule": "each frame uses a different composition and angle", "base_shot_type": "m

登入查看完整提示詞

登入解鎖