Generate images with Google's Nano Banana Gemini 2.5 Flash Image model via OpenRouter API using text prompts.
# Nano Banana OpenRouter Skill
Generate images using Google's **Nano Banana** (Gemini 2.5 Flash Image) models via **OpenRouter API**.
## Configuration
Add the following to your `openclaw.json` (or set the env var `OPENROUTER_API_KEY`):
```json
{
"skills": {
"entries": {
"nano-banana-openrouter": {
"enabled": true,
"config": {
"apiKey": "sk-or-v1-..."
},
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-..."
}
}
}
}
}
```
## Usage
**Tool Name:** `nano_banana_generate`
**Examples:**
- "Generate a cyberpunk dragon using Nano Banana."
- "Draw a landscape in 16:9 aspect ratio."
- "Use the preview model to generate a logo for a coffee shop."
## Models
- **Default:** `google/gemini-2.0-flash-exp:free` (Free tier, robust text/code, experimental image support)
- **High Quality:** `google/gemini-2.5-flash-image-preview` (The specific "Nano Banana" image model, paid/credits required)
## Notes
- OpenRouter requires the `HTTP-Referer` and `X-Title` headers (included in this skill).
- `modalities: ["image"]` is sent automatically.
don't have the plugin yet? install it then click "run inline in claude" again.