Head-top Diorama Prompt for Nano Banana Pro Agent V1.2 (Afro/Cloud Amusement Park)

A complex, multi-part system prompt for Nano Banana Pro Agent V1.2 to create a 'Head-top Diorama' image, where the character's head or hat serves as the base for a miniature, fantastical scene. This specific example focuses on an afro/voluminous hair interpreted as a cloud amusement park. This is a duplicate of tweet 1998335885985669354.

Prompt

2,525 charactersOpen Astorie studio ↗
You are a **"Miniature Set Designer and Portrait Photographer" who constructs fantastical worlds.** Deeply analyze the content of the input text provided at the end, and generate an image of an **elaborate and narrative "Head-top Diorama"** developed using the person's head (or hat) as a base.

At that time, place the character from the attached reference image as the **"wearer (the base person)"** of the diorama. Maintain the consistency of the character's facial features and identity (ID) while depicting a natural expression suited to the weight and atmosphere of the scene on their head.

▼ Instructions for Generation Process (Thinking Process)
Utilize Nano Banana Pro's inference capability and execute the following thinking process before drawing:

1. **[Step 1: World Scaling]**: Select elements to condense the theme specified in the input text (city, forest, factory, etc.) to hat size.
2. **[Step 2: Boundary Fusion]**: Logically design how the brim of the hat or the hair physically connects to the ground of the diorama (coastline, cliff, floor).
3. **[Step 3: Micro-Lighting]**: Simulate what kind of shadows (Cast Shadow) the light sources within the diorama (streetlights, bonfire, magical light) cast on the wearer's face.

▼ Design and Style Specification
* **Composition/Layout**: Bust-up portrait. The camera is slightly high-angle (俯瞰) to show details on the head. Shallow depth of field (DoF), focusing on the diorama part and blurring the background to emphasize the miniature feel (Tilt-Shift effect).
* **Taste**: Photorealistic, macro photography, fantastical, cinematic lighting.
* **Text**: Accurately draw text symbolizing the theme (e.g., "{argument name="sign text 1" default="跡地"}", "{argument name="sign text 2" default="Cafe"}", etc.) in natural fonts, in Japanese or English, on signs or markers placed within the diorama.
* **Character**: The reference character is either quietly closing their eyes or posing as if looking up to feel the world on their head.

▼ Input Text
[Afro/Voluminous Hair × Cloud Amusement Park]
**Goal:** Interpret the hair's "fluffiness" as "clouds" to construct a fancy world.
> **Input Text:**
> キャラクターのボリュームのある髪(アフロやカーリーヘア)を、ピンク色の夕焼け雲に見立てた「天空の遊園地」のジオラマ。
> * **風景:** 髪の雲の上に、観覧車やメリーゴーランド、ジェットコースターのレールが埋まっている。風船がいくつも空へ飛んでいく。
> * **光の演出:** 夕日が逆光となり、髪(雲)の輪郭が金色に輝く(リムライト)。遊園地のイルミネーションがキラキラと点灯し始めている。
> * **テキスト:** 遊園地の入り口ゲートに、ポップなフォントで**「{argument name="amusement park name" default="DREAM LAND"}」**という立体文字が浮かんでいる。
> * **詳細:** 髪の柔らかい質感が、綿菓子や雲のようなエーテル状の物質に見える.
From sourceView source ↗

About

The source requires reference images. Confirm input files before use.

More to explore

Social Media Post - Red Velvet Leotard Portrait Comparison
Profile / Avatar - Purple Sweater Portrait Comparison
Profile / Avatar - Nano Banana Pro Lake Portrait Comparison
Profile / Avatar - Cute Star Night Light Portrait Comparison
Social Media Post - Nano Banana Pro Silver Metallic Gown Portrait
Profile / Avatar - Early 2000s Internet Cafe Webcam Couple

Profile / Avatar - Early 2000s Internet Cafe Webcam Couple

Fav PromptNano Banana Pro
A candid couple photo of the Woman's Look and Man's Look in the Setting, engaged in the Activity. Render it as the Camera Type, with soft focus, smeared facial detail, color noise, dim uneven indoor light and low-resolution compression. Use awkward webcam framing from a desk-level angle, with a slight tilt and casually cropped shoulders or arms. Include the Era Props naturally in the booth. The couple should look genuinely close and relaxed, with real skin, messy everyday hair and no heavy retouching. Use Face Reference for the matching faces when applicable. Make it feel like an unedited early-2000s personal photo from an old Facebook album or MSN chat, not a modern image with a retro filter. No date stamp, text, watermark, modern electronics or current fashion. Output in the Format.

Use these details:
Woman's Look: "{argument name="woman's look" default="Young woman in her early 20s with straight black hair, a casual graphic T-shirt and a loose zip hoodie"}"
Man's Look: "{argument name="man's look" default="Young man in his early 20s with a thin mustache, tousled hair and a worn casual sweatshirt"}"
Setting: "{argument name="setting" default="A cramped internet cafe booth with a laminate desk, worn carpet, beige walls and bulky CRT monitors"}"
Activity: "{argument name="activity" default="Leaning close together and looking into the webcam during an online chat"}"
Camera Type: "{argument name="camera type" default="Low-resolution built-in webcam capture from a chunky desktop computer"}"
Era Props: [{argument name="era props" default="Wired earphones, A burned CD in a paper sleeve, A chunky desktop keyboard"}]
Face Reference: "{argument name="face reference" default="Use attached photos for the matching person's face when provided; if none are attached, invent both faces"}"
Format: "{argument name="format" default="4:3 landscape"}"

Find your next direction

ASTORIE CANVAS

Take the idea further.

Bring your prompts, references and results together in one creative space.

Open your canvas ↗