Nano Banana Desktop

⚠️ Work in Progress - v1.0.0 Release

A desktop image editing utility for Linux that leverages Google's Gemini Flash 2.5 Images (Nano Banana) for AI-powered image transformations.

This is an early-stage experimental project that provides a functional desktop interface for applying AI-powered image transformations using Google's Gemini API. The basic implementation has been validated and works reliably for image-to-image transformations using prompt templates.

Project Status

This v1.0.0 release represents a minimal viable product that demonstrates:

• Working Gemini API integration for image transformations
• Desktop GUI with tab-based workflow
• Prompt template library for common edits
• Basic image versioning and backup

Future Plans:

• Support for additional API providers (Claude, Replicate, etc.)
• Enhanced editing capabilities
• More sophisticated prompt management
• Performance optimizations
• Expanded template library

Note: This project was developed using AI-assisted coding (Claude Code) and is released as-is for experimental use.

Features

• Tab-based interface for managing multiple images
• AI-powered image editing using Gemini's image-to-image capabilities
• Prewritten prompt templates for common edits
• Custom prompt support
• Version control with automatic backup
• Iterative editing workflow
• Multiple aspect ratio support (1:1, 16:9, 21:9, 9:16)
• Secure API key storage

Screenshots

Requirements

• Python 3.10 or higher
• Linux with KDE/Qt support
• Google Gemini API key

Installation

Option 1: Install from .deb Package (Recommended for Ubuntu)

Download and install the latest .deb package:

# Download the latest release from GitHub
# Visit: https://github.com/danielrosehill/Nano-Banana-Desktop/releases

# Install the package
sudo dpkg -i nano-banana-desktop_1.0.0_amd64.deb

# Fix dependencies if needed
sudo apt-get install -f

After installation, run from terminal:

nano-banana-desktop

Or search for "Nano Banana Desktop" in your application menu.

Updating:

./update-deb.sh  # Auto-downloads and installs latest release
# OR
./update-deb.sh path/to/new-package.deb  # Install specific package

Option 2: Quick Start from Source

1. Clone the repository:

git clone https://github.com/danielrosehill/Nano-Banana-Desktop.git
cd Nano-Banana-Desktop

2. Run the setup script:

./setup-venv.sh

3. Run the application:

./run.sh

Manual Installation

If you prefer to set up manually:

cd code
uv venv .venv
source .venv/bin/activate
uv pip install -e ".[dev]"

Usage

Coming soon!

Building

Build Standalone Executable

To build a standalone PyInstaller executable:

./build.sh

This will create a self-contained executable at ./dist/nano-banana-desktop that can be run without needing Python installed.

Build .deb Package

To build a Debian package for Ubuntu:

./build-deb.sh

This creates a .deb package at ./dist/nano-banana-desktop_1.0.0_amd64.deb that can be distributed and installed on Ubuntu systems.

The .deb package:

• Installs to /opt/nano-banana-desktop
• Creates launcher command nano-banana-desktop
• Adds desktop entry to application menu
• Manages virtual environment automatically
• Handles dependencies through apt

Development

This project uses:

• PySide6 for the GUI
• google-genai for Gemini API integration
• Pillow for image processing
• keyring for secure API key storage

Project Structure

Nano-Banana-Desktop/
├── code/                    # Main codebase
│   ├── nano_banana/        # Python package
│   │   ├── ui/             # GUI components
│   │   ├── core/           # Core functionality and API integration
│   │   └── utils/          # Utility functions (includes prompt loader)
│   ├── prompts/            # Prompt template library
│   │   ├── artistic-effects/
│   │   ├── color-adjustments/
│   │   ├── background-editing/
│   │   └── ... (14 categories total)
│   └── tests/              # Test suite
├── docs/                   # API documentation
├── dev-spec/               # Development specifications
├── setup-venv.sh           # Virtual environment setup script
└── run.sh                  # Application runner script

Prompt Templates

The application includes 14 categories of pre-written prompt templates:

• Artifact Removal
• Artistic Effects (comic book, impressionist, pop art, etc.)
• Atmosphere
• Avatars
• Background Editing
• Color Adjustments
• Enhancements
• Expression Edits
• For Fun
• Graphics
• Style Transfer
• Subject Separation
• Text Editing

License

TBD

Author

Daniel Rosehill