25 Nano Banana Pro prompt ideas (copy & use)
This post turns scattered “prompt inspiration” into 25 reusable prompt directions. Pick a scene, replace the placeholders, and iterate—this is the fastest way to build your own nanobanana prompt library.

Before you copy: 3 rules to increase hit rate
- Start with output specs: aspect ratio, resolution, text allowed or not.
- Lock non-negotiables: identity, product shape, logo position, exact wording.
- Two-pass workflow: first pass for layout; second pass for small text and polish.
25 prompt directions (copy & edit)
Replace placeholders in
[ ]. Keep brand and technical names like Nano Banana Pro and nanobanana2 unchanged.
- Sticker pack (LINE-style 4×6)
Turn the reference person into a cute LINE Friends-style chibi sticker pack. 4×6 grid, each tile is one expression + short Chinese handwritten phrase. Constraints: do not copy the original image; clean outlines; text fully readable. Output 16:9, 4K.
- Photo → chibi cartoon (big head, thick outline)
Convert all people in the photo into a cute chibi illustration: big head small body, rounded proportions, thick clean black outline, minimal facial features, slight blush, flat colors with almost no gradient. Keep everything else unchanged: background, lighting, perspective, composition, colors, clarity.
- Photoreal selfie brief (structured constraints)
Photoreal candid selfie: [PERSON] + [POSE] + [OUTFIT/JEWELRY] + [LOCATION] + [LIGHT]. Natural skin texture, correct fingers, face in sharp focus, mild background depth-of-field. Output 3:4, high-res.
- 3×3 cinematic grid (same identity, different angles)
Create a 3×3 cinematic photo grid. Same character identity across all tiles (face, hair, outfit unchanged), each tile uses a different shot type (close-up, over-shoulder, profile, etc.). Unified color grade + subtle film grain. Output 1:1, 4K.
- Wide quote card (portrait 1/3 + text 2/3)
Wide quote card: brown background, light-gold serif typography. Portrait on left (1/3), quote on right (2/3). Quote: “[QUOTE]”, author: “—[AUTHOR]”. Constraints: text perfectly readable, no typos, no distortion. Output 16:9, 4K.
- Vintage “time-mismatch” ink painting humor
Traditional Chinese gongbi ink painting on aged xuan paper: a Tang-dynasty noblewoman holding a modern hair dryer; servants using a modern smartphone. Pine trees and rocks in background; calligraphy + red seal. Soft mineral pigments. Output 16:9, high-res.
- 3D cutaway infographic (English labels)
High-tech 3D cutaway infographic of [OBJECT]. Semi-disassembled cutaway with floating parts and callout lines. Clear English labels + short function descriptions. Clean background, highly organized layout, like an engineering blueprint. Output 16:9, 4K, no watermark.
- Fitness teaching diagram (steps + arrows)
Create a professional step-by-step diagram for [EXERCISE]. At least 5–6 actions, each split into 2–4 key phases. Use arrows to show motion paths; highlight alignment/safety cues. Chinese titles and labels, fully readable. Output 9:16, high-res.
- Black-and-white fashion couple shoot
Black-and-white fashion editorial. Same couple, same outfits, different poses. Minimal gray seamless background. Soft studio lighting (Rembrandt/butterfly). High-contrast monochrome conversion. No text. Output high-res.
- Illustration → photoreal “humanized”
Convert [ILLUSTRATION] into photoreal studio portrait while keeping pose and composition. Natural pores, realistic fabric folds, no plastic skin. Output 1:1, high-res.
- Vocabulary scene labeling (3-line labels)
Draw a detailed [SCENE] and label objects with a 3-line format: line1 English, line2 IPA, line3 Chinese. Labels must not block the subject and must be fully readable. Output high-res.
- Hand-drawn recipe infographic (Z-layout)
Hand-drawn recipe infographic, 9:16. Beige paper texture, watercolor sketch + ink outlines. Z-shaped reading flow, 3–5 steps, each includes illustration + short text + chef tip. Final dish illustration at the bottom. Output high-res, no watermark.
- 4×4 “decade transformation” grid
Make a 4×4 grid showing the same person across decades from [START] to [END]. Each tile matches its era (outfit, hair, accessories, color grade, film look). Identity consistent. Output 1:1, 4K.
- Poem watercolor with pinyin
Watercolor illustration expressing the mood of “[POEM]”. Add pinyin (with tones) above every Chinese character. Text must be perfectly readable, no deformation. Output 3:4, high-res.
- Character sheet (outfit layers + props + materials)
Create a character sheet: full-body character in the center, surrounded by outfit layers, expressions, core props, material close-ups. Hand-drawn arrows connect parts. Parchment background. Notes in Chinese:English labels. Output 4K.
- OOTD collage (left person, right item list)
Create an OOTD collage: street-style portrait on the left; a clean white product list on the right (item photo + brand + Chinese item name + price). Minimal editorial layout, modern sans-serif font. Output 4K.
- Hand-drawn city travel map (handwriting + seal)
Hand-drawn travel map of [CITY] in watercolor + ink on aged parchment. Isometric micro-landmarks. All labels are handwritten calligraphy; include curved text along a river path; add a red seal stamp in a corner. Output 16:9, high-res, no garbled text.
- Three-panel facial emotion strip (top/middle/bottom)
Based on the reference person, generate a horizontal three-panel strip (top/middle/bottom). Identity unchanged. Top: eyes+mouth close-up with tears. Middle: warm smile. Bottom: profile close-up, playful pout. Output 16:9, high-res.
- Futuristic face analysis infographic
Make a futuristic face analysis infographic: subtle scan grid overlay + a red vertical laser line. Five metrics with labels and percentages. Bottom center: “Estimated Age: [NUMBER]”. Modern sans-serif type, fully readable. Output high-res.
- Kids literacy poster (sticker labels + soft 3D)
Create a kids literacy poster “([SCENE] vocabulary)”. A4 vertical, 2.5D picture-book look with gentle DOF. Add sticker-like labels with 3 lines: pinyin / Chinese / English, with callout lines. Output 8K, crisp readable text.
- 4-panel comic (conflict + warm punchline)
4-panel comic: main character + angel on left shoulder + devil on right shoulder + a third small animal. Angel says study, devil says relax, animal suggests eating first. Desk room scene. Cute fable style. Chinese. Constraints: distinct character designs, warm funny ending. Output high-res.
- E-commerce long page (mobile readable)
Create a Chinese e-commerce long page for [PRODUCT]: hero visual, 3–5 key selling points, material/process close-ups, use cases, size table, after-sales, CTA section. Clear hierarchy, mobile readable, no garbled text. Output 1:4 or 1:8, high-res.
- Group scene with balanced weights (no overlap)
Cinematic group photo in one continuous scene with [N] people. Faces clearly recognizable, natural scale, balanced visual weight, subtle interactions, no blocking. No text, no logos, no watermark. Output 16:9, high-res.
- Ultra-wide panorama poster (4:1)
Ultra-wide concept panorama: high detail density but organized composition; strong depth layering; no text, no watermark. Output 4:1, 4K.
- “Chat bubble as table” micro-photography concept
Photoreal top-down macro shot: a light-green chat bubble becomes a long dining table. Two realistic tiny humans sit at both ends eating. Bubble contains only the text “[TEXT]”, with a gray timestamp above. Background is a dense messaging-themed line pattern. Output high-res.