nanobanana-skill

Image generation from text prompts uses AI models to create visuals based on natural language descriptions. This Skill leverages Google Gemini API to generate images in multiple aspect ratios and resolutions—square, landscape, or ultra-wide—by sending your prompt and configuration options to the model, which returns a ready-to-save image file.

Yes, image editing with natural language instructions allows you to modify existing images by describing the changes you want. This Skill accepts an image file and edit prompt, sends both to Gemini API, and returns an edited version without requiring manual design software.

Gemini API supports multiple aspect ratios—portrait, landscape, and ultra-wide—with resolution options including 1K, 2K, and 4K. This Skill lets you specify --size and --resolution flags to control output dimensions and quality before generation or editing.

Yes, this Skill supports model selection via the --model flag: gemini-3-pro-image-preview prioritizes higher quality output, while gemini-2.5-flash-image prioritizes faster generation. Choose based on whether quality or speed matters more for your workflow.

To use Gemini image generation, you need a valid GEMINI_API_KEY, Python 3 with installed dependencies (google-genai, Pillow, python-dotenv, httpx[socks]), and the nanobanana.py executable. Store your API key in a .env file and ensure all packages are installed before running.

Yes, this Skill is designed for scripted image generation and editing. It accepts command-line options (--size, --model, --resolution) and saves output as image files, making it suitable for automation, batch processing, or integration into larger creative pipelines.