gemini-imagegen

Generate and edit images via Gemini API with text prompts, supports text-to-image, editing, style transfer, and multi-reference composition. Always save images as .jpg due to JPEG default format.

• Generate images from text prompts
• Edit existing images with text prompts
• Apply style transfers and composition from multiple references
• Supports real-time search grounding

• Uses gemini-3-pro-image-preview (1K-4K) as default model
• Supports 1K, 2K, 4K resolutions
• Available aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9
• Multi-turn refinement with chat-based iterative editing
• Can combine up to 14 reference images
• Returns images in JPEG format by default
• Supports Google Search grounding for real-time data
• Images include SynthID watermarks

1. 1.Set environment variable GEMINI_API_KEY
2. 2.Use generate_content with model='gemini-3-pro-image-preview' and specify image_config for resolution/aspect ratio
3. 3.Pass existing images with text prompts for editing
4. 4.Use chat interface for multi-turn refinement with iterative prompts
5. 5.Convert to PNG only if needed via PIL after saving as JPEG
6. 6.Always save generated images with .jpg extension