Index of every page on this site: llms.txt
All skills
inference-sh avatar

/nano-banana-2

@c895a47
by inference.shinference-sh/skills819 stars
108
Run it yourself

Generate images with Google Gemini Nano Banana 2.1, Google's newest image model, via inference.sh CLI. Capabilities: text-to-image, image editing, multi-image input (up to 14 images), Google Search grounding, up to 4K. Triggers: nano banana 2.1, nano banana 2, nanobanana 2, gemini nano banana 2.1, gemini 3.1 flash image, google image generation

Use this Skill: https://skilld.dev/gh/inference-sh/skills/nano-banana-2

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈90 tokens for metadata: the name and description. ≈1.1k when used: this file.

Description uses 4.5% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈362 characters, or 4.5%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Install the belt CLI skill: npx skills add belt-sh/cli

Nano Banana 2.1 - Gemini Nano Banana 2.1

Generate images with Google Gemini Nano Banana 2.1, Google's newest image model, via inference.sh CLI.

Quick Start

Requires inference.sh CLI (belt). Install instructions

belt login

belt app run google/gemini-nano-banana-2-1 --input '{"prompt": "a banana in space, photorealistic"}'

Examples

Basic Text-to-Image

belt app run google/gemini-nano-banana-2-1 --input '{
  "prompt": "A futuristic cityscape at sunset with flying cars"
}'

Multiple Images

belt app run google/gemini-nano-banana-2-1 --input '{
  "prompt": "Minimalist logo design for a coffee shop",
  "num_images": 4
}'

Custom Aspect Ratio

belt app run google/gemini-nano-banana-2-1 --input '{
  "prompt": "Panoramic mountain landscape with northern lights",
  "aspect_ratio": "16:9"
}'

Image Editing (with input images)

belt app run google/gemini-nano-banana-2-1 --input '{
  "prompt": "Add a rainbow in the sky",
  "images": ["https://example.com/landscape.jpg"]
}'

High Resolution (4K)

belt app run google/gemini-nano-banana-2-1 --input '{
  "prompt": "Detailed illustration of a medieval castle",
  "resolution": "4K"
}'

With Google Search Grounding

belt app run google/gemini-nano-banana-2-1 --input '{
  "prompt": "Current weather in Tokyo visualized as an artistic scene",
  "enable_google_search": true
}'

Input Options

Parameter Type Description
prompt string Required. What to generate or change
images array Input images for editing (up to 14). Supported: JPEG, PNG, WebP, HEIC, HEIF
num_images integer Number of images to generate
aspect_ratio string Output ratio: "1:1", "16:9", "9:16", "4:3", "3:4", "21:9", "4:1", "1:4", "8:1", "1:8", "auto"
resolution string "1K", "2K", "4K" (default: 1K)
output_format string Output format for images
enable_google_search boolean Enable real-time info grounding (weather, news, etc.)

Output

Field Type Description
images array The generated or edited images
description string Text description or response from the model

Prompt Tips

Styles: photorealistic, illustration, watercolor, oil painting, digital art, anime, 3D render

Composition: close-up, wide shot, aerial view, macro, portrait, landscape

Lighting: natural light, studio lighting, golden hour, dramatic shadows, neon

Details: add specific details about textures, colors, mood, atmosphere

Sample Workflow

# 1. Generate sample input to see all options
belt app sample google/gemini-nano-banana-2-1 --save input.json

# 2. Edit the prompt
# 3. Run
belt app run google/gemini-nano-banana-2-1 --input input.json

Python SDK

from inferencesh import inference

client = inference()

# Basic generation
result = client.run({
    "app": "google/gemini-nano-banana-2-1@55q9sjpt",
    "input": {
        "prompt": "A banana in space, photorealistic"
    }
})
print(result["output"])

# Stream live updates
for update in client.run({
    "app": "google/gemini-nano-banana-2-1@55q9sjpt",
    "input": {
        "prompt": "A futuristic cityscape at sunset"
    }
}, stream=True):
    if update.get("progress"):
        print(f"progress: {update['progress']}%")
    if update.get("output"):
        print(f"output: {update['output']}")

Related Skills

# Original Nano Banana (Gemini 3 Pro Image)
npx skills add inference-sh/skills@nano-banana

# Full platform skill (all apps)
npx skills add inference-sh/skills@infsh-cli

# All image generation models
npx skills add inference-sh/skills@ai-image-generation

Browse all image apps: belt app list --category image

Documentation

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warning1d5 checks · Risk SAFE
  • Gen Agent Trust Hub2d

    The skill facilitates image generation using the Google Gemini Nano Banana 2.1 model via the author's belt CLI and Python SDK. No malicious patterns or security risks were identified.

  • Socket1d

    No alerts

  • Snyk2d

    Risk: MEDIUM · 1 issue

  • Runlayer8mo

    1 file scanned · No issues

  • ZeroLeaks7mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at c895a47. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 4 days ago
What it can do
Runs commands
All 1 allowed tools
Bash(belt *)
  • gemini
  • image-generation
  • text-to-image
  • image-editing
  • google
  • inference-sh
  • belt
  • bash

README badge

README badge for inference-sh/skills/nano-banana-2

Generates images with Google Gemini 3.1 Flash Image Preview via the inference.sh CLI, supporting text-to-image, image editing with up to 14 input images, custom aspect ratios, and real-time Google Search grounding. Use the belt command-line tool to run the model with parameters like prompt, resolution (up to 4K), and num_images.

Generated from the current SKILL.md.

What model does this skill use?
Google Gemini 3.1 Flash Image Preview (also called Nano Banana 2), accessed via the inference.sh CLI.
Can I edit existing images with this skill?
Yes. Pass up to 14 images in the `images` parameter along with a prompt describing the edits you want.
What image formats does this support?
Input images can be JPEG, PNG, or WebP. Output format is configurable via the `output_format` parameter.
Does this work with real-time information?
Yes, if you set `enable_google_search: true`, the model can ground generations in current weather, news, and other live data.
What aspect ratios and resolutions are available?
Aspect ratios include 1:1, 16:9, 9:16, 4:3, 3:4, and auto. Resolutions range from 1K (default) to 2K and 4K.

Generated from the current SKILL.md. These answers refresh after source changes.