Image Generation
Kaji can generate images using AI via the Nanobanana integration, powered by Google's Gemini image generation models. Describe what you want and Kaji will produce the image directly in your session.
What Kaji can do
- Generate new images from text descriptions
- Edit existing images using natural language
- Combine multiple input images
- Control aspect ratio, resolution, and style
- Generate multiple variations at once
Example prompts
Generate a professional banner image for our "Kaji AI Assistant" product launch.
Clean, modern style with a dark background and purple accents.
Create an icon for a data analytics dashboard — abstract, minimal, blue tones
Generate a diagram showing how AI agents connect to tools in a company
Edit this logo to have a white background instead of transparent
Create 3 variations of a hero image for our website homepage
Output options
| Option | Values |
|---|---|
| Aspect ratio | 1:1, 16:9, 9:16, 4:3, 3:4, and more |
| Resolution | High (default), 2K, 4K |
| Model | Flash (fast), Pro (high quality) |
| Count | 1–4 images |
Tips for business users
- Be descriptive: include style, color palette, mood, and composition
- For best results, mention the intended use ("for a presentation", "as a website banner")
- Ask for "flat design" or "minimalist" for clean professional visuals
- You can ask for edits to a generated image by describing what to change
Configuration
Nanobanana uses Google's Gemini image generation API.
| Variable | Description |
|---|---|
GEMINI_API_KEY | Google Gemini API key |
IMAGE_OUTPUT_DIR | Local directory where generated images are saved |