Chrome ExtensionNano Banana 2 Prompt Generator
Generate optimized prompts for Google's Nano Banana 2 (Gemini 3.1 Flash Image) — 2× faster than Nano Banana Pro with web search grounding, controllable Thinking mode, multilingual text rendering, and up to 4K resolution. Built around Google's official Creative Director-style prompting and the latest 2026 best practices.
Be specific: describe who/what is in the image with details
Quick Palettes
Generated Prompt
Fill in the form and click "Generate" to create an optimized Nano Banana 2 prompt.
Tip: Be as specific as possible with your subject description. Instead of "a dog," try "a golden retriever puppy with a red bandana."
Nano Banana 2 Tips
- • Write like a Creative Director — full sentences, never tag soup ("dog, park, 4k")
- • Nano Banana 2 holds up to 4-5 characters and 14 reference objects with consistency — name them all
- • Turn Web Grounding ON for real-world subjects (locations, monuments, species) — not for people
- • Keep Thinking Mode OFF by default; switch ON for complex infographics or grounding + spatial reasoning
- • Up to 4K resolution and extreme aspect ratios (4:1, 1:4) supported — name the resolution tier you want
- • Multilingual text rendering and translation works inline — request the target language explicitly
Nano Banana 2 Prompt Templates
Copy-ready image briefs built around what this model actually rewards. Swap the [BRACKETED] parts for your own subject and settings, then paste straight into the tool.
Multi-Character Scene
CharacterNano Banana 2 holds consistency across several characters at once.
[NUMBER] characters in [SETTING]. Character A: [FULL DESCRIPTION — build, hair, clothing with materials]. Character B: [FULL DESCRIPTION]. Character C: [FULL DESCRIPTION]. They are [WHAT THEY ARE DOING / HOW THEY RELATE SPATIALLY]. [LIGHTING], [CAMERA ANGLE]. Each character stays visually distinct and consistent with the description given.
Grounded Real-World Subject
PhotographyWeb-search grounding makes real places, products and landmarks accurate.
[REAL SUBJECT — specific landmark, product model, or location by name] photographed [FRAMING] in [CONDITIONS, e.g. late afternoon light, light rain]. Render the actual architectural and material details accurately. [FOREGROUND ELEMENT]. [LENS/PERSPECTIVE]. Photorealistic, true-to-life proportions and colour.
Multilingual Text Graphic
TypographyMultilingual text rendering with an explicit layout.
A [GRAPHIC TYPE] with the headline "[TEXT]" in [LANGUAGE/SCRIPT] across the top in [FONT STYLE], and "[SECOND TEXT]" in [SECOND LANGUAGE] beneath it in a smaller face. [VISUAL ELEMENT] below the text block. [BACKGROUND]. Both scripts rendered correctly and legibly, balanced hierarchy, [ASPECT RATIO].
Extreme Aspect Ratio Banner
LayoutUses the 4:1 / 8:1 ratios most models cannot handle.
An ultra-wide banner. [SUBJECT] positioned at [POSITION, e.g. the far left third], with [SECONDARY ELEMENT] balancing the right. Wide empty space in the centre for overlay text. [BACKGROUND] extending seamlessly across the full width. [LIGHTING]. Composition designed to read at 8:1 — nothing important near the extreme edges.
4K Detail Study
PhotographyHigh-resolution work where fine texture matters.
Extreme close-up of [SUBJECT], filling the frame. Every [TEXTURE DETAIL, e.g. fibre, pore, scratch, water droplet] is resolved sharply. [LIGHTING, e.g. raking side light picking out surface relief]. Shallow depth of field with the focal plane on [WHERE]. Macro photography, 4K detail, natural colour.
Thinking-Mode Composition
ScenesControllable Thinking mode for logically constrained scenes.
[SCENE WITH A LOGICAL OR PHYSICAL CONSTRAINT — e.g. a room where every reflective surface shows a consistent reflection of the same window]. Verify before rendering that [THE CONSTRAINT, e.g. shadow directions all agree with a single light source at upper right, and the item count is exactly N]. [STYLE], [LIGHTING], [ASPECT RATIO].
How to Use the Prompt Generator
Pick Mode & Describe Your Vision
Choose Generate, Edit, or Multi-Image Blend. Pick a use case category, set resolution tier (512px → 4K), toggle Thinking and Web Grounding, and describe your subject. For real-world subjects, leave Web Grounding ON for factual accuracy on locations, monuments, and species.
Generate Your Prompt
Click Generate. The tool drops you into a Creative Director-style brief — natural-language sentences, concrete materials and lighting, explicit color palette mapping, multilingual text directives if needed, and a preserve list when editing.
Copy & Create
Paste into the Gemini app, Google AI Studio, the Gemini API (model: gemini-3.1-flash-image-preview), or Vertex AI. For 4K hero shots, draft at 512px first to iterate cheaply, then upscale the winner.
Frequently Asked Questions
What is Nano Banana 2?
Nano Banana 2 (officially Gemini 3.1 Flash Image) is Google's latest image generation model, launched February 26, 2026. It combines Gemini 3 Pro's world knowledge and reasoning with Flash-tier speed — roughly 2× faster than Nano Banana Pro at a fraction of the cost. It is the default image model across the Gemini app, Google Search, AI Mode, Flow, and Google Ads.
How does this prompt generator work?
You fill in structured inputs — subject, scene, style, composition, lighting, color palette, generation mode, and intended use. Our AI then assembles them into a Nano Banana 2-optimized prompt: natural-language sentences (not tag soup), camera and material specifics, multilingual text directives, and explicit preservation lists for edits.
What changed from Nano Banana Pro to Nano Banana 2?
Nano Banana 2 is roughly 2× faster than Pro at ~25% the cost while keeping ~95% of Pro's capabilities. New: a controllable Thinking mode (toggle reasoning ON/OFF), exclusive 512px tier for cheap drafts, extreme aspect ratios (4:1, 8:1, 1:4, 1:8), web search grounding for real-world subjects, character consistency for up to 4 characters, and reference blending across up to 14 images per workflow.
What are Generate, Edit, and Multi-Image Blend modes?
Generate creates an image from text. Edit modifies a single uploaded photo via natural-language instructions — fill in the Preserve field (face, pose, lighting, brand text) so the model doesn't drift. Multi-Image Blend composites up to 14 references — describe each by index ("Image 1: base scene; Image 2: jacket reference") and how they should interact.
When should I turn Thinking Mode on?
Keep Thinking OFF by default for speed. Turn it ON only when you hit nonsensical outputs, are building complex infographics with dense data, or need to combine web grounding with spatial reasoning (e.g., "lay out a labeled map of Italy with these 5 cities at correct relative positions").
When should I enable Web Search Grounding?
Turn web grounding ON for real-world subjects where factual accuracy matters: famous locations, monuments, biological species, branded products, or specific historical events. Web grounding cannot be used for people. Leave it OFF for stylized work, fictional scenes, or speed-critical drafts.
What resolution and aspect ratios does Nano Banana 2 support?
Resolution tiers: 512px (3-8s, exclusive to NB2 for cheap drafts), 1K (default, 5-15s), 2K (10-25s), 4K (15-40s). Standard aspect ratios: 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16. Exclusive extreme ratios for banners and tall panels: 4:1, 8:1, 1:4, 1:8.
How accurate is text rendering in Nano Banana 2?
Text rendering and inline multilingual translation are first-class features. For best results: wrap literal text in quotes or ALL CAPS, specify font style and color, state placement explicitly, and for tricky brand names, spell them letter-by-letter. Localization works inline — you can request the same image with text in multiple languages.
Why natural language instead of keyword tags?
Nano Banana 2 inherits Gemini's reasoning DNA — it understands full sentences far better than comma-separated keyword lists. Google's official guide explicitly recommends Creative Director-style briefs over "tag soup" like "dog, park, 4k, realistic." Our generator enforces this and replaces vague praise with concrete visual facts.
Is this tool free to use?
Yes. You get 1 free prompt generation per day with no signup required. For unlimited access, sign up for a Promptslove membership which includes all AI tools and 20,000+ premium prompts.
Where can I use the generated prompts?
The prompts are tuned for Nano Banana 2 via the Gemini app, Google AI Studio, the Gemini API (model: "gemini-3.1-flash-image-preview"), and Vertex AI. They also produce strong results on Nano Banana Pro (Gemini 3 Pro), GPT Image 2, Midjourney, FLUX.2, and Seedream 5.
Want Unlimited AI Prompt Generation?
Get unlimited access to all AI tools, 20,000+ premium prompts, courses, and resources designed to maximize your creative output.