Display Name
Nano Banana
Category
Agent Skills
Primary Link
https://github.com/Ibrahim-3d/nano-banana-claude-plugin
Author Name
Ibrahim
Author Link
https://github.com/Ibrahim-3d
License
MIT
Description
A Claude Code plugin that brings Google Gemini's native image generation API into Claude Code through a single /genimage slash command, an autonomous AI agent, and 8 Python scripts. Supports text-to-image, text-guided image editing (inpainting, add/remove objects, background replacement), style transfer, multi-image composition (up to 14 references), 4K resolution generation, Google Search grounded generation, and multi-turn iterative editing. All editing is text-guided with no visual UI — Gemini semantically understands natural language region descriptions.
Validate Claims
Install the plugin, set up a Gemini API key in scripts/.env, then run /genimage a watercolor lighthouse during a storm. The agent selects texttoimage.py, runs it via the Gemini API, and saves the output image to disk.
Specific Task
Install the plugin and ask Claude to edit an existing image: /genimage edit myphoto.png - replace the background with a tropical beach
Specific Prompt
/genimage a photorealistic golden retriever puppy on a rainy window ledge, warm bokeh background, 16:9
Additional Comments
- "Nano Banana" is Google's internal codename for Gemini's native image generation model
- Uses gemini-2.5-flash-image (fast) and gemini-3-pro-image-preview (4K, search grounding)
- Makes external calls to the Google Gemini API (requires user's own API key from aistudio.google.com)
- Requires Python 3.9+, google-genai, python-dotenv, Pillow packages
- MIT licensed
Recommendation Checklist
Display Name
Nano Banana
Category
Agent Skills
Primary Link
https://github.com/Ibrahim-3d/nano-banana-claude-plugin
Author Name
Ibrahim
Author Link
https://github.com/Ibrahim-3d
License
MIT
Description
A Claude Code plugin that brings Google Gemini's native image generation API into Claude Code through a single
/genimageslash command, an autonomous AI agent, and 8 Python scripts. Supports text-to-image, text-guided image editing (inpainting, add/remove objects, background replacement), style transfer, multi-image composition (up to 14 references), 4K resolution generation, Google Search grounded generation, and multi-turn iterative editing. All editing is text-guided with no visual UI — Gemini semantically understands natural language region descriptions.Validate Claims
Install the plugin, set up a Gemini API key in scripts/.env, then run
/genimage a watercolor lighthouse during a storm. The agent selects texttoimage.py, runs it via the Gemini API, and saves the output image to disk.Specific Task
Install the plugin and ask Claude to edit an existing image:
/genimage edit myphoto.png - replace the background with a tropical beachSpecific Prompt
/genimage a photorealistic golden retriever puppy on a rainy window ledge, warm bokeh background, 16:9Additional Comments
Recommendation Checklist