nanobanana-skill

Image generation with Gemini API uses prompt-based requests to create visuals. This Skill generates images by accepting text prompts, aspect ratios, and model parameters, then returns generated images saved as output files. You need a GEMINI_API_KEY, Python 3 with google-genai and Pillow installed, and the nanobanana tool.

Both image editing and generation are supported. This Skill accepts input images for editing tasks alongside prompt-based generation. You provide an existing image file, a prompt describing edits, and output specifications like aspect ratio and resolution.

Prerequisites include configuring GEMINI_API_KEY in ~/.nanobanana.env or as an environment variable, installing Python 3 with google-genai, Pillow, and python-dotenv packages, and having the nanobanana.py executable available in your path.

Yes, aspect ratios and custom resolutions are configurable parameters. You specify these alongside your prompt and model choice when generating or editing images, giving control over output dimensions.

Generated and edited images are output as PNG files by default. The Skill handles format conversion and file saving through Pillow, so you receive ready-to-use image files without manual format handling.

Model selection is a configurable parameter in this Skill. Different Gemini models may produce varying results; you specify the model alongside your prompt, aspect ratio, and resolution for each generation or editing task.