Nano Banana V2 - AI Image Creation Tool#

English | 简体中文

Overview#

Nano Banana V2 is a powerful AI image generation and smart splitting tool based on the AceData Nano Banana series models. It provides a complete web interface for image generation, editing, splitting, and management.

Core Features#

🎨 Image Generation

• Multiple Models: Nano Banana Pro, Nano Banana 2
• Multiple Resolutions: 1:1 (Square), 16:9 (Landscape), 9:16 (Portrait), 4:3 (Standard), 3:4 (Portrait)
• Multiple Qualities: 1K (1024px), 2K (2048px), 4K (4096px)
• Image-to-Image: Support up to 6 reference images for editing

✂️ Smart Splitting

• Multiple Grids: Support 2/4/6/9 grid splitting
• Smart Adaptation: Automatically detect aspect ratio and choose optimal splitting method

🖼️ Gallery Management

• Waterfall Layout: Automatically adapt to different image ratios
• Thumbnail Generation: Automatic thumbnail creation
• Batch Download: Support ZIP packaging
• File Management: One-click open file directory

🌐 Multi-Language Support

• 简体中文 (zh)
• English (en)
• 繁體中文 (zh-TW)
• 日本語 (ja)
• 한국어 (ko)

🎨 User Experience

• Dual Themes: Light/Dark theme switching
• Real-time Preview: Image preview, split preview
• Modern UI: Frosted glass effect, smooth animations, waterfall layout
• AI Prompt Generation: Integrated LLM for intelligent prompt generation

---

Screenshots#

Main Interface#

---

Quick Start#

1. Install Dependencies#

bash

cd ~/.openclaw/workspace/skills/nano-banana-v2
npm install

2. Start Service#

bash

npm start

3. Access Application#

Open browser at http://localhost:2688

---

Configuration#

Get API Key#

First-time use requires API Key configuration:

1. Visit: https://share.acedata.cloud/r/1uN88BrUTQ
2. Register and get API Key
3. Fill in API Key in settings page

Configuration Items#

ItemRequiredDescriptionAPI Key✅For image generationPlatform Token❌For image-to-image featureLLM API Key❌For AI prompt generationSave Path❌Image save directory (Default: Desktop/banana2)

---

Usage#

Generate Image#

1. Enter prompt in the bottom editor
2. Select model, resolution, quality
3. Select splitting method (optional)
4. Click generate button
5. Wait for completion

Image-to-Image#

1. Click upload button to select images (max 6)
2. Enter description
3. Click generate

AI Prompt Generation#

1. Click AI button to open prompt generation window
2. Enter simple description
3. Click generate, AI will create detailed prompt

View Works#

• Click image to view full size
• View split image list
• Download, delete works

---

Tech Stack#

• Backend: Node.js + Express + SQLite
• Frontend: HTML5 + CSS3 + JavaScript (jQuery)
• Image Processing: Sharp
• API: AceData Nano Banana API

---

Directory Structure#

text

nano-banana-v2/
├── server/
│   ├── index.js              # Main server
│   └── services/             # Service classes
│       ├── database.js       # Database service
│       ├── request.js        # Request handling
│       ├── task.js           # Task processing
│       └── upload.js         # Upload handling
├── public/
│   ├── index.html            # Main page
│   ├── list.html             # List page
│   ├── set.html              # Settings page
│   ├── components/           # UI components
│   ├── css/                  # Styles
│   ├── js/                   # Application logic
│   └── lan/                  # Language files
├── config/
│   ├── set.json              # Main config
│   ├── system.prompt         # AI system prompt
│   └── cut_*.prompt          # Split prompts
├── database/
│   ├── 1.sql                 # Database init
│   └── 2.sql                 # Database migration
├── package.json
├── README.md
├── LICENSE
└── CHANGELOG.md

---

API Endpoints#

MethodPathDescriptionGET/api/get_setGet configPOST/api/generateSubmit generation taskPOST/api/poll/:idPoll task statusPOST/api/uploadUpload imageGET/api/worksGet works listGET/api/work/:idGet single workPOST/api/tasks/addAdd task manuallyPOST/api/admin/delete/:idDelete workPOST/api/open-folderOpen folderPOST/api/generate-promptAI generate promptPOST/api/shutdownShutdown service

---

Changelog#

v1.1.0 (2026-03-30)#

• Added Korean (한국어) support
• Completed multi-language translation for all components
• Optimized dynamic image path mapping
• Cleaned up redundant API endpoints
• Fixed known issues

v1.0.0 (2026-03-26)#

• Initial release

---

Open Source Info#

• License: MIT License
• Repository: https://github.com/xiyunnet/banana2
• Issues: https://github.com/xiyunnet/banana2/issues
• Homepage: https://banana2.zjhn.com

---

Contact#

• Author: Xiao Zhu
• Email: zhuxi0906@gmail.com
• WeChat: jakeycis