nano-image-generator

The Nano Image Generator is a versatile automation tool designed to create high-quality visual assets directly from text prompts. Leveraging the capabilities of the Nano Banana Pro (Gemini 3 Pro Preview) model, it serves developers, designers, and content creators who need to rapidly prototype or produce graphics such as app icons, marketing banners, UI elements, and website hero images. By integrating this skill into your workflow, you can automate the generation of visuals, ensuring consistent and professional branding across your digital projects. This tool is particularly effective for generating illustrative content, diagrams, and social media assets without requiring deep graphic design expertise.

• Powered by Gemini 3 Pro Preview (Nano Banana Pro) for high-fidelity image synthesis from natural language descriptions.

• Supports multiple aspect ratios including square (1:1), portrait, and landscape orientations (e.g., 16:9, 9:16) for tailored asset creation.

• Adjustable output resolution settings (1K, 2K, 4K) to match specific project requirements.

• Configurable API integration requiring standard environment variables (APIKEY, BASEURL), with built-in fallback to GEMINI_API_KEY.

• Command-line interface optimized for local execution, allowing for quick integration into automated build scripts or development pipelines.

• The script requires a valid API key from the Gemini API provider to function; configure this in the .env file within the project root.

• When crafting prompts, specify the desired style (e.g., flat design, 3D rendered, photorealistic) and context (e.g., mobile app icon) to achieve the most relevant output.

• Output paths should be managed carefully; recommended directories include /assets/icons/, /public/images/, or /marketing/ for organized asset storage.

• While ideal for icons and banners, users should avoid requesting complex embedded text, as generative models perform best with visual composition rather than typography.

• Use the provided flag system (--aspect, --size) to ensure that generated assets are pre-formatted for their intended media, reducing the need for post-processing.