Generate images via RunComfy's google/nano-banana-2/text-to-image endpoint.
npx skills add https://github.com/runcomfy-com/skills --skill nano-banana-2-runcomfy-com
Please help me install this Agent Skill. Skill: nano-banana-2 Source: https://github.com/runcomfy-com/skills/tree/main/nano-banana-2 Command: npx skills add https://github.com/runcomfy-com/skills --skill nano-banana-2-runcomfy-com
It helps users generate higher-quality images with Google Nano Banana 2 on RunComfy by using the model’s proven prompting patterns instead of generic text-to-image instructions.
Use the nano-banana-2 skill to create an image by requesting an image generation with the phrase “nano-banana-2” followed by your full subject and style description, including any exact text to render in the image.
High-intent search queries and answers about installing and using this skill.
To generate images with Google Nano Banana 2, use a subject-first prompting strategy by declaring the subject before style, and trigger the RunComfy CLI endpoint google/nano-banana-2/text-to-image with your prompt parameters. This ensures higher-quality image generation compared to generic text-to-image instructions.
The best way to prompt in-image typography for Nano Banana 2 is to enclose the exact text you want rendered in quotes within your prompt. Use a subject-first declarative phrasing strategy to ensure the model accurately generates the specified text inside the image.
Yes, you can batch generate image variants by configuring the num_images parameter in the JSON input schema. You can also select social-native aspect ratios, choose resolution tiers from 0.5K to 4K, and set output formats for your final drafts.
Yes, Nano Banana 2 supports optional web grounding by enabling the enable_web_search parameter in the JSON input schema. This allows the text-to-image model to fetch real-world references for entities and events during image generation.
To refine and maintain consistent image generation results, use stable refinement via the seed parameter in the JSON input schema. Adjusting safety_tolerance and limit_generations also helps control output variation and generation limits across drafts.
agentspace-soGenerate images from natural-language prompts via the RunComfy Model API.
prime-skillsGenerate polished still images from natural-language prompts via RunComfy.
GeeveGeorgeSearch and verify web images to ground AI image generation in real-world visuals.
doany-aiGenerate image drafts with Google Nano Banana 2 via the RunComfy CLI.
