Randolph's Not So Nano Banana

Quick Start Workflow

1. Enter a descriptive text prompt
2. Choose optional settings (style, aspect ratio, size, negative prompts)
3. Click "Work the Image"
4. Wait 5-15 seconds for your image
5. Use output controls to refine, share, or export

Core Features

• Prompt: Describe what you want to see. Be specific for best results.
• Generate: Click "Work the Image" to generate at your selected resolution.
• Draft: Quick 1K preview on NanoBanana Flash — fast & cheap. Upscale later if you like the result.
• Style Preset: 20 artistic styles (photorealistic, anime, poster, logo design, infographic, etc).
• Aspect Ratio: 14 options. Auto-detects from uploaded images. Flash supports extreme ratios (1:8, 8:1); ChatGPT models max out at 21:9.
• Image Size: 1K (fast), 2K (balanced), or 4K (maximum quality — Gemini, and experimental on ChatGPT Image 2.5; Mini tops out at 1K).
• Quality (ChatGPT): Auto/Low/Medium/High — the main cost lever on ChatGPT models.
• Transparent Background (ChatGPT): Alpha PNG cutouts for stickers and product shots.
• Batch Generation: Create 1-4 variations simultaneously.
• Prompt Helper: Browse 325+ curated prompt blocks across 16 categories. Click to build your prompt.
• Upload Images: Add up to 14 reference images — large photos are auto-compressed to fit, keeping full resolution wherever possible. Images are auto-named (e.g., "sunset_01"). Type @ in your prompt for autocomplete.
• Enhance Text: Improved accuracy for text, labels, and typography in images.
• Thinking Mode: AI reasoning for higher quality and more complex outputs (NanoBanana Flash only).
• Controls-Only Mode: Apply professional settings without changing image content.

Advanced Features

• Real Time Data: Use Google Search for current events, weather, stock prices, etc.
• @Mention Autocomplete: Type @ to reference uploaded images by name with dropdown suggestions.
• AI Model: Five models across two providers — NanoBanana Flash/Pro (Gemini) and ChatGPT Image 2.5 Flare/Sunburst/Mini (OpenAI). Pick per job; see each option's strengths in the dropdown.
• Conversation History: Track every iteration. Click any past generation to restore and branch from it.
• Prompt Enhancement: Get 3 AI-powered prompt suggestions with detailed improvements.
• Thought Signatures: View encrypted signatures and comprehensive generation metadata.
• Cross-Provider Iteration: Refine any image with any model — Gemini and ChatGPT chains interoperate automatically.

Professional Controls

• 3D Camera Control: Interactive 3D widget with azimuth (horizontal rotation), elevation (vertical angle), and distance sliders. Camera description auto-added to prompt.
• Camera Angles: Eye level, low angle, high angle, bird's eye, worm's eye, dutch angle, over-the-shoulder.
• Lighting: Natural, studio, dramatic, soft, hard, backlit, golden hour, blue hour, neon.
• Depth of Field: Shallow (blurred background), medium, deep, extreme bokeh.
• Focus: Sharp/crisp, soft focus (dreamy), tilt-shift (miniature), selective.
• Color Grading: Cinematic, vibrant, muted, warm, cool, vintage, film noir, high contrast.

Output Controls

• Share: Open native share menu (mobile) or copy to clipboard (desktop).
• Export: Choose format (PNG, JPEG, WebP) and quality.
• Download: Instantly save as PNG with smart filename.
• New Image: Reset everything and start fresh.

Iteration Mode

After generating an image, the app enters Iteration Mode:

• Generated image becomes the base for refinement
• Enter a new prompt to modify the current image
• Previous image automatically included in next request
• Click "New Image" to exit and start fresh
• AI thinking process shown in real-time during generation

Keyboard Shortcuts

• Enter (in prompt) - Generate image
• Alt + Shift + N - Reset interface (New Image)
• Ctrl/Cmd + H - Open Prompt Helper
• @ (in prompt) - Reference uploaded images
• Escape - Close modals