nanobanana-pro-fallback
by yazelin
Image & Video GenerationBrief
Nano Banana Pro with auto model fallback — generate/edit images via Gemini Image API.
Description
What This Skill Does
Generates and edits images using Google's Gemini Image API, with an automatic fallback chain across multiple Gemini models. Accepts text prompts for new images or existing images for editing, at 1K, 2K, or 4K resolution. Supports compositing up to 14 input images into a single output.
The automatic model fallback means generation continues even when a specific Gemini model is unavailable or rate-limited.
When to Use It
- Generate a product mockup image from a text description
- Edit a photo by describing the changes in natural language
- Combine multiple reference images into one scene
- Produce marketing visuals at different output resolutions
- Script repeated image generation with timestamped filenames
Example Workflow
User asks: generate a 2K image of a tropical beach at sunset, save it as beach.png
- 1
Run generate_image.py via uv with the prompt and filename arguments
- 2
Set --resolution 2K for higher quality output
- 3
Script tries gemini-2.5-flash-image first, falls back to gemini-2.0-flash-exp-image-generation if needed
- 4
Script prints a MEDIA: line with the saved file path for OpenClaw to attach
beach.png saved at 2K resolution, auto-attached to the chat via the MEDIA: output line
Requirements
- GEMINI_API_KEY environment variable (from Google AI Studio)
Installation
clawhub install yazelin/nanobanana-pro-fallbackRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.