Nano Banana Pro — OpenClaw Skill

What This Skill Does

The Nano Banana Pro skill gives your OpenClaw agent the ability to generate new images from text prompts, edit existing images, and compose multiple images together using Google's Gemini 3 Pro Image model. It supports three operations: text-to-image generation, single-image editing with natural language instructions, and multi-image composition with up to 14 input images.

Images can be generated at three resolution tiers -- 1K (default, fastest), 2K (balanced), and 4K (highest quality). The bundled Python script handles the Gemini API interaction and outputs a MEDIA: line that OpenClaw auto-attaches on supported chat providers like WhatsApp, Telegram, and Discord. Use timestamps in filenames for organized output.

This is the go-to skill for any visual content creation workflow: generating social media graphics, editing product photos, creating illustrations for blog posts, compositing screenshots into a single overview image, or producing quick mockups for design discussions.

Example Prompts

Generate an image of a cozy coffee shop interior with warm lighting and plants

Remove the background from this product photo and make it transparent

Create a 4K banner image for our blog post about AI agents in the workplace

Combine these three screenshots into a single side-by-side comparison image

Edit this photo to make the sky more dramatic with sunset colors

Generate a minimalist logo concept for a tech startup called "Lumina"

Take this wireframe sketch and turn it into a polished UI mockup

Create an isometric illustration of a server room for our infrastructure docs

Requirements

Dependencies: uv (Python package runner) and a GEMINI_API_KEY.

Install uv via Homebrew: brew install uv
API key: Get a Gemini API key from ai.google.dev
Environment variable: Set GEMINI_API_KEY or configure in OpenClaw config under skills."nano-banana-pro".env.GEMINI_API_KEY

Setup on KiwiClaw

Add your Gemini API key in the KiwiClaw dashboard settings. The skill is pre-installed with uv and ready to generate images immediately. On Standard plans, the Gemini key can be pooled for included usage.

Setup Self-Hosted

Install uv: brew install uv
Get a Gemini API key from ai.google.dev
Set GEMINI_API_KEY in your environment or OpenClaw config
The skill activates when you ask your agent to generate or edit images

Related Skills

Video Frames -- extract frames from video to use as input for image editing
Xurl -- post generated images to X/Twitter
WaCLI -- send generated images via WhatsApp
Sag -- pair generated images with TTS narration for multimedia content

FAQ

What image resolutions does Nano Banana Pro support?

Nano Banana Pro supports three resolution tiers: 1K (default, fastest), 2K (balanced quality), and 4K (highest quality, slower). Specify the resolution with the --resolution flag when generating images.

Can Nano Banana Pro edit existing images?

Yes. Pass an existing image with the -i flag along with edit instructions in the --prompt. For example: --prompt "remove the background" -i photo.png. The skill supports editing single images or compositing up to 14 images together.

What API key is needed?

A GEMINI_API_KEY environment variable is required. You can get one from Google AI Studio. Alternatively, set it in your OpenClaw config under skills."nano-banana-pro".env.GEMINI_API_KEY.

Can Nano Banana Pro combine multiple images?

Yes. Multi-image composition supports up to 14 input images. Pass multiple -i flags with different image paths and a prompt describing how to combine them into one scene.