nanobanana-skill

To generate AI images using Gemini models, you run the provided Python script with a text prompt to produce and save high-quality visuals. This approach wraps Gemini image generation models to streamline single-prompt outputs for marketing assets and mockups.

You can combine multiple reference images into one cohesive output using multi-reference composition. By providing several input images alongside your prompt, the tool merges them while maintaining a consistent style across the generated visual.

Gemini image generation supports live web search grounding to ensure up-to-date visuals and factual accuracy. This grounded generation feature leverages live Google search data during the creation process to produce relevant, context-aware image outputs.

You need a GEMINI_API_KEY configured in your environment to run multi-reference image editing workflows. The tool also requires dependencies from requirements.txt and the executable Python script present in your working directory to process prompts and save outputs.

The best way to create product mockups with adjustable aspect ratios is using a flexible Gemini-powered wrapper that lets you select models and resolutions. It saves the generated images alongside optional text notes and metadata for auditing.

If generated images are not saving, check the debugging logs produced by the executable script. The tool outputs saved images, optional text, metadata, and logs specifically to help trace file path issues or API failures during batch generation.