Nano Banana Pro 图像生成

What This Skill Does

Nano Banana Pro (Nano Banana 2) is a cutting-edge image generation and editing agent skill for OpenClaw, powered by Google's latest high-fidelity AI models. This skill enables users to transform text prompts into stunning, photorealistic visuals or modify existing images through natural language instructions. Whether you need custom digital art, complex image compositing, or high-resolution photo generation, Nano Banana Pro handles the heavy lifting by automatically detecting your intent based on the presence of source images. It supports advanced features like text rendering within images, various aspect ratios, and configurable output resolutions (1K, 2K, 4K).

Installation

To integrate this capability into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/hexiaochun/sutui-nano-banana-pro

Ensure your agent environment has sufficient credits, as each generation task consumes 60 credits. Once installed, the tool becomes immediately available for your agent to trigger via the submit_task interface.

Use Cases

Nano Banana Pro is highly versatile and serves several key domains:

• Creative Design: Generating unique assets for presentations, social media, or marketing materials.
• Image Editing: Seamlessly merging multiple source images or altering subjects within a frame using text descriptions.
• Content Creation: Creating photorealistic imagery for articles or blog posts by referencing current web trends via the enable_web_search parameter.
• Prototyping: Quickly visualizing concepts with specific aspect ratios, such as 21:9 for cinematic previews or 9:16 for mobile content.

Example Prompts

1. "Generate a 4K photorealistic image of a futuristic cyberpunk city skyline at night with neon lights reflecting on wet streets."
2. "Take this image of my dog and place him in a snowy mountain landscape, make it look like a professional wildlife photograph."
3. "Create an image of a vintage coffee shop sign with the text 'OpenClaw Cafe' written in a retro serif font."

Tips & Limitations

To maximize the quality of your output, provide descriptive prompts that specify camera angles, lighting conditions, and stylistic requirements (e.g., 'cinematic lighting', 'macro photography'). The model excels at text rendering, so ensure your instructions for specific text are clear.

Limitations: The model is optimized for high-end generation and consumes 60 credits per request. While it supports multi-image input for editing, ensure your image_urls are accessible and correctly formatted. Always review the get_task output, as complex edits may require iterative refinements. Enable enable_web_search only when you require up-to-date visual references from the web to improve accuracy for contemporary topics.