nano-banana-edit

Edit images with Google Nano Banana 2 via the RunComfy CLI.

12|2|Updated Aug 12, 2026
One-click install
npx skills add https://github.com/genmedia-labs/skills --skill nano-banana-edit-genmedia-labs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: nano-banana-edit
Source: https://github.com/genmedia-labs/skills/tree/main/nano-banana-edit
Command: npx skills add https://github.com/genmedia-labs/skills --skill nano-banana-edit-genmedia-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @runcomfy/cli.

What problem does it solve? Editing images at scale—swapping backgrounds, localizing changes, or batch-editing up to 20 images while preserving subject identity—normally requires manual design work or complex ComfyUI pipelines. This Skill routes those edits to Google's Nano Banana 2 image-to-image endpoint through the RunComfy CLI with a single command. ## Core Features & Use Cases - Identity-preserving edits: Swap backgrounds or clothing while keeping the subject's identity, pose, and framing unchanged. - Batch editing: Pass 1–20 input images per call with locked aspect ratio and resolution for consistent SKU galleries and ad variants. - Spatial localization: Target edits with language like "the left object only" or "background only" for precise, scoped changes. - Routing guidance: Built-in decision table for when to use GPT Image 2 edit, Flux Kontext, or Nano Banana 2 text-to-image instead. - Use Case: An e-commerce team edits 20 product photos in one batch, replacing each background with a consistent studio sweep while preserving the product exactly. ## Quick Start Ask the agent to edit an image with Nano Banana, for example: "Use nano banana edit to swap the background of this product photo to a neon cyberpunk street while keeping the subject unchanged."

Frequently Asked Questions about nano-banana-edit

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I edit an image with Nano Banana on RunComfy?▼

Install the RunComfy CLI with npm i -g @runcomfy/cli, run runcomfy login, then call runcomfy run google/nano-banana-2/edit with a JSON input containing your prompt and image_urls array. Results download to the directory set by --output-dir.

Nano Banana Edit vs Flux Kontext vs GPT Image 2 for image editing?▼

Use Nano Banana Edit for identity-preserving background or clothing swaps and batches up to 20 images. Use Flux Kontext for single-reference precise local edits, and GPT Image 2 edit for multilingual text inside images.

How many images can Nano Banana Edit process in one call?▼

The endpoint accepts 1 to 20 input images per call, with the first treated as primary and the rest as auxiliary references. It returns 1 to 4 output images per call.

Why does my Nano Banana edit change the subject's face or brand?▼

The prompt is missing preservation goals. Lead with an explicit statement like "Keep the subject identity, pose, and clothing unchanged" before describing the change, since unstated elements get subtly rewritten.

What are the limitations of the Nano Banana 2 edit endpoint?▼

It accepts at most 20 inputs and 4 outputs per call, long compound prompts drift and should be split into multiple passes, and web search adds latency and cost. Multilingual in-image text edits are better handled by GPT Image 2 edit.