Master prompt architecture, camera framing, and lighting techniques for Nano Banana 2. Explore 2,020 curated and deduplicated Nano Banana 2 prompts categorized across 16 creative industries.
Nano Banana 2 is built around advanced world knowledge, reliable subject consistency, precise instruction following, improved text rendering and localization, versatile image generation and editing capabilities, and rapid iteration cycles. Crafting cohesive scene descriptions with clear visual direction enables the model to translate complex concepts into refined images.
Rather than using disjointed keyword lists or generic quality phrases like "hyperrealistic, 8k, masterpiece", provide clear, descriptive sentences that establish the subject, surroundings, and visual context naturally.
Writing articulate descriptions gives the model clear context for subject features, background elements, and overall composition.
A practical prompting workflow organizes scene elements logically: Subject → Action & Pose → Environment → Camera Framing & Lens Perspective → Lighting Character → Textures & Materials.
Organizing your prompt in a coherent sequence clarifies spatial relationships and helps maintain stylistic harmony throughout the image.
Describe tangible visual attributes: specific materials (e.g. matte brushed aluminum, weathered concrete), identifiable light sources, and defined framing boundaries rather than abstract superlatives.
Concrete descriptions provide actionable visual cues that guide surface textures, highlights, and scene details effectively.
A production-grade prompt is composed of distinct building blocks. While not every generation requires all fourteen elements, selecting the right blocks for your specific genre gives complete creative control.
The central entity: persona demographics, age, clothing, posture, or physical product geometry and material properties.
What the subject is actively doing. Avoid passive stiffness; specify interaction with physical objects or bodily posture.
Architectural or landscape context, background depth layers, floor textures, and props that ground the scene.
Framing strategy: rule of thirds, centered balance, intentional negative space for typography, or leading lines.
Where the lens is positioned in 3D space: eye-level, low-angle looking up, high-angle, top-down flat lay, or Dutch tilt.
Controls optical perspective, compression, and field of view (e.g. 24mm wide environmental vs. 85mm portrait compression).
Source, quality, and direction of light: soft window daylight, hard sun, studio softbox, rim light, or low-key chiaroscuro.
Aperture guidance: shallow f/1.8 focus isolating the subject against creamy bokeh, or deep f/11 focus for architectural clarity.
Color harmony and film stock characteristics: muted earthy tones, high-contrast monochrome, or warm amber and cool teal contrast.
Micro-textures: porous stone, brushed matte titanium, frosted glass refraction, crisp linen weave, or moist dewy droplets.
Verbatim lettering enclosed in exact quotation marks with font type, placement, and size specified.
Concise elimination of unwanted visual artifacts, artificial smoothing, or accidental text clutter.
The most common reason for unpredictable AI image output is an underspecified subject. Compare how changing simple phrases into concrete visual descriptors transforms consistency in Nano Banana 2.
| Domain | Vague Subject (Unpredictable) | Production-Ready Subject (Deterministic) | Why It Works |
|---|---|---|---|
| Portrait | Vague "A beautiful woman in a cafe" | Ready "A woman in her early 30s with shoulder-length wavy chestnut hair, light freckles over her nose, wearing an oversized beige ribbed knit sweater, sipping an espresso" | Anchors demographics, specific hair geometry, wardrobe texture, and interactive physical action. |
| Product | Vague "A bottle of luxury perfume on a table" | Ready "A heavy rectangular flacon of amber perfume made of clear faceted glass, capped with a brushed gold cylindrical stopper, resting on a slab of honed black travertine" | Specifies packaging geometry, glass refraction, metallic finish, and exact material grounding. |
| Automotive | Vague "A black sports car on the road" | Ready "A matte satin frozen-black 2026 sports coupe with flared aerodynamic wheel arches, dark bronze multi-spoke wheels, and red brake calipers, carving an asphalt mountain hairpin" | Defines exact paint finish (satin matte vs gloss), aerodynamic bodykit, wheel spec, and driving posture. |
| Food | Vague "A gourmet burger with fries" | Ready "A thick smashed beef patty with a charred crispy crust, melted aged cheddar draping over the edges, crisp butterhead lettuce, and glossy brioche bun sprinkled with toasted sesame seeds" | Specifies patty cooking technique, cheese viscosity, salad freshness, and bread gloss. |
Using photographic terminology serves as effective prompting guidance for field of view, spatial framing, perspective lines, and emotional scale in Nano Banana 2.
Fills the frame with a specific micro-feature: an eye, watch dial, lip texture, or ring facet.
Frames head and neck. Focuses on facial expressions, makeup, skin realism, and emotion.
From chest up. Balances subject intimacy with immediate shoulder attire and collar details.
From waist up. Natural conversational framing, ideal for hand gestures and workspace interactions.
From knees up ("cowboy shot"). Captures posture, holster/belt accessories, and room architecture.
Complete subject from head to toe. Ensures full outfit, footwear, and ground contact are visible.
Subject is integrated into the larger landscape or room. Establishes scale and environmental context.
Vast landscape or cityscape where the subject is tiny or silhouetted against monumental scale.
Direct human eye height. Neutral, intimate, and documentary-like without forced perspective bias.
Shooting upward from below chest or knee height. Bestows power, grandeur, confidence, and architectural height.
Shooting downward toward the subject. Emphasizes floor geography, subject vulnerability, or spatial arrangement.
Directly perpendicular to the surface. Essential for culinary styling, desk layouts, and product flat lays.
Angled 45 degrees between front and profile. Reveals depth, jawline contours, and three-dimensional volume.
Canted horizon line. Injects kinetic urgency, psychological tension, dynamic action, or avant-garde style.
Expands spatial volume, exaggerates leading lines, and delivers grand architectural sweep. Keep human faces away from extreme edges to avoid barrel stretching.
Replicates natural human optical perception. The quintessential choice for street photography, editorial lifestyle, and natural reportage.
Flattens perspective, compresses foreground and background, eliminates facial elongation, and produces buttery smooth background separation.
A common source of visual confusion in multi-angle prompting is changing the viewpoint around a subject while inadvertently keeping the original background static.
Maintaining Spatial Coherence: When prompting for a new camera angle—such as moving from front to back—explicitly update the background and lighting direction to match the new perspective. If the camera looks at the subject from behind, the objects originally in front of the subject now appear in the background, while original background elements drop out of frame.
"View of the sports car from behind. Same shiny glass skyscrapers directly behind the car, same frontal headlights gleaming."
Result: Visual contradiction between the rear car body and unchanged front background.
"Rear 3/4 view of the sports car. The street now stretches into the distance behind it toward an open intersection; the glass skyscrapers are behind the camera and their reflections gleam in the rear windshield. Afternoon sun hits from the opposite side, casting forward shadows."
Result: Visually cohesive composition with aligned perspective and lighting cues.
Specifying the light source, quality, and direction provides clear guidance for surface highlights, shadow character, and overall mood in Nano Banana 2.
Gentle, natural illumination filtered through sheer fabric or north-facing glass. Soft shadow gradients, perfect for natural editorial portraits and interior morning scenes.
Low-angle warm sunlight (approx. 2800K–3200K) just after sunrise or before sunset. Long dramatic shadows, warm edge rimming, and luminous atmospheric glow.
Cool ambient twilight (approx. 7000K–9000K) after the sun has dipped below the horizon. Deep cobalt skies contrasting with warm artificial window or street lanterns.
High-noon or crisp unobstructed sun. Sharp-edged, high-contrast shadows, punchy color saturation, and high specular reflections on metal, glass, and water.
Large diffused key light used in commercial studio photography. Flawless highlight roll-off, controlled fill, and round catchlights in the subject's eyes.
Hard raw flash mounted directly above the lens. Produces gritty Y2K editorial aesthetics, sharp drop-shadows on walls, and hyper-vibrant pop colors.
Strong light source positioned behind the subject facing the lens. Creates a bright halo outlining hair, shoulders, or product contours to separate from dark backgrounds.
Beams of sunlight cutting through airborne particles, dust, mist, or smoke. Imparts dramatic cinematic majesty in cathedrals, forests, or industrial halls.
Dramatic contrast between concentrated highlights and deep velvety shadows. Inspired by Caravaggio; ideal for noir mysteries, fine-art portraits, and luxury packaging.
Homogeneous, bright, shadow-free lighting against white or pastel backdrops. Essential for skincare, medical technology, e-commerce, and cheerful corporate visuals.
Opposing saturated neon sources (e.g. electric cyan key light and magenta rim light). Quintessential for cyberpunk, nightlife editorial, and modern music aesthetics.
Visible in-scene illumination sources such as desk lamps, neon signs, candles, or computer displays that cast authentic localized luminescence onto adjacent surfaces.
How elements are positioned within the frame determines viewer attention and hierarchy. Combine compositional structures with optical depth controls.
Leaving expansive unencumbered areas (e.g. clean monolithic walls, empty sky, or calm water) around a concentrated subject. Crucial for commercial advertising, hero banners, and editorial typography placement.
Placing an out-of-focus object in the immediate foreground (e.g. doorframe, tree branch, window mullion, or shoulder) to add multi-layered depth and voyeuristic realism.
Utilizing architectural corridors, road centerlines, bridges, or railway tracks that draw the viewer's gaze from the bottom edges directly into the focal center.
Use "shallow depth of field, f/1.4, creamy circular bokeh" when subject isolation is paramount. Use "deep depth of field, f/11, tack-sharp edge-to-edge clarity" for complex street tableaus, architecture, and diagrams.
When using image-to-image or multi-reference pipelines in Nano Banana 2, assign explicit operational roles to each reference image so the model understands what to inherit and what to generate afresh.
Lock the human subject's facial bones, eye shape, nose bridge, ear geometry, and distinctive facial features. State clearly:
Lock manufacturing contours, packaging proportions, logo typography, and material textures while completely altering the staging:
Borrow color temperature, film grain, and contrast curve without copying scene contents:
When modifying an existing generation, do not regenerate the entire image from scratch. Use the structured Preserve vs. Change formula to maintain pixel integrity where desired while executing targeted revisions.
Separate your edit prompt into two explicit sections: what to freeze and what to modify.
Nano Banana 2 has superior typographical rendering capabilities. To achieve legible, pristine in-scene lettering, follow strict formatting rules.
Always wrap desired text in exact straight double quotation marks. Specify the exact casing (UPPERCASE or Title Case):
Describe font weight, style, and finish (e.g. geometric sans-serif, high-contrast Didot serif, embossed gold foil, neon glow).
AI models often populate background posters, books, and signage with gibberish. Explicitly restrict stray lettering in the negative constraint:
Identify and resolve the most frequent prompting errors that lead to compromised image quality, spatial anomalies, or unexpected outputs.
Practical questions regarding Nano Banana 2 prompt engineering and deployment.
Explore 2,020 curated and deduplicated Nano Banana 2 prompts. Filter by industry category, toggle reference image requirements, or execute multi-keyword searches. Click any card to copy the full prompt.
