sc-gemini-imagegen

Text-to-image generation with Gemini converts your written descriptions into images at 1K-4K resolution. Set your GEMINI_API_KEY environment variable, then use the included script with a prompt like `python scripts/generate_image.py --prompt "A sunset over mountains"` to create images directly from text.

Image editing with Gemini supports multi-turn refinement, letting you iteratively modify and improve visuals in a single workflow. You can adjust compositions, regenerate sections, and refine outputs across multiple rounds without restarting.

Gemini image generation offers selectable image sizes and aspect ratios through configurable parameters. The gemini-3-pro-image-preview model handles these options, letting you specify dimensions suited to your design needs like logos, stickers, or product mockups.

Composition from multiple reference images is supported, enabling you to blend or combine separate visual inputs into complex outputs like logos, stickers, and product mockups within a single generation task.

Image generation requires google-genai for API access and Pillow for image processing and manipulation. Both dependencies are included with the Skill and handle generation, editing, and output formatting automatically.

Generated images return as inline image data alongside text output, making them immediately ready to save to disk or pass into downstream image processing and editing workflows without format conversion.