gemini-imagen 0.6.0

A comprehensive Python wrapper for Google Gemini's image generation and analysis capabilities with S3 and LangSmith integration

gemini-imagen

A comprehensive Python library and CLI for Google Gemini's image generation and analysis capabilities.

📚 For Python library usage, see LIBRARY.md 🚀 For advanced features, see ADVANCED_USAGE.md 🤝 For contributing, see CONTRIBUTING.md

Features

• 🎨 Text-to-Image Generation - Create images from text prompts
• 📐 Aspect Ratio Control - Custom aspect ratios (16:9, 1:1, 9:16, etc.)
• 🏷️ Labeled Input Images - Reference images by name in prompts
• 📸 Multiple Output Images - Save same image to multiple locations
• 💬 Image Analysis - Get detailed text descriptions of images
• ☁️ S3 Integration - Seamless AWS S3 upload/download with URL logging
• 📈 LangSmith Tracing - Full observability for debugging and monitoring
• 🔒 Safety Settings - Configurable content filtering thresholds
• 🖥️ CLI Tool - Powerful command-line interface for all operations
• 🔄 Type-Safe - Full type hints with Pydantic validation

Installation

Basic Installation

pip install gemini-imagen

With S3 Support

pip install gemini-imagen[s3]

From Source

git clone https://github.com/aviadr1/gemini-imagen.git
cd gemini-imagen
pip install -e ".[dev,s3]"

Quick Start

CLI Usage

# Set up your API key
export GOOGLE_API_KEY="your-api-key-here"

# Or save it in config
imagen keys set google YOUR_API_KEY

# Generate an image
imagen generate "a serene Japanese garden with cherry blossoms" -o garden.png

# Analyze an image
imagen analyze photo.jpg

# Edit an image
imagen edit "make it sunset" -i original.jpg -o edited.png

# Upload to S3
imagen upload local.png s3://my-bucket/remote.png

Python Library

For detailed Python API documentation, see LIBRARY.md.

Quick example:

from gemini_imagen import GeminiImageGenerator

generator = GeminiImageGenerator()

# Generate an image
result = await generator.generate(
    prompt="A serene Japanese garden with cherry blossoms",
    output_images=["garden.png"]
)

print(f"Image saved to: {result.image_location}")

CLI Commands

The CLI provides comprehensive image generation and management capabilities:

Command	Description	Example
generate	Generate images from text prompts	imagen generate "a cat" -o cat.png
analyze	Analyze and describe images	imagen analyze image.jpg
edit	Edit images using reference images	imagen edit "make it brighter" -i photo.jpg -o out.png
upload	Upload images to S3	imagen upload local.png s3://bucket/remote.png
download	Download images from S3	imagen download s3://bucket/image.png local.png
keys	Manage API keys	imagen keys set google YOUR_KEY
config	Manage configuration	imagen config set default_model gemini-2.0-flash-exp
models	List and manage models	imagen models list

Common CLI Options

# Generate with options
imagen generate "prompt" -o output.png \
  --temperature 0.8 \
  --aspect-ratio 16:9 \
  --safety-setting preset:relaxed \
  --trace \
  --json

# Use input images
imagen generate "blend these styles" \
  -i style.jpg --label "Style:" \
  -i composition.jpg --label "Composition:" \
  -o result.png

# Pipe input
echo "a sunset" | imagen generate -o sunset.png
cat prompt.txt | imagen generate -o output.png

Python Library Examples

For comprehensive Python API documentation, examples, and integration patterns, see LIBRARY.md.

Here are a few quick examples:

Text-to-Image Generation

result = await generator.generate(
    prompt="A futuristic cityscape at sunset with flying cars",
    output_images=["cityscape.png"],
    aspect_ratio="16:9",
    temperature=0.8
)

Image Analysis

result = await generator.generate(
    prompt="Describe this image in detail",
    input_images=["photo.jpg"],
    output_text=True
)
print(result.text)

With Safety Settings

from gemini_imagen import SafetySetting, HarmCategory, HarmBlockThreshold

result = await generator.generate(
    prompt="A tasteful artistic photo",
    output_images=["output.png"],
    safety_settings=[
        SafetySetting(
            category=HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT,
            threshold=HarmBlockThreshold.BLOCK_ONLY_HIGH
        )
    ]
)

For more examples including S3 integration, LangSmith tracing, batch processing, and web framework integration, see LIBRARY.md.

Configuration

Environment Variables

# Required
export GOOGLE_API_KEY=your_google_api_key

# Optional - for S3 features
export GV_AWS_ACCESS_KEY_ID=your_aws_access_key
export GV_AWS_SECRET_ACCESS_KEY=your_aws_secret_key
export GV_AWS_STORAGE_BUCKET_NAME=your-bucket-name

# Optional - for LangSmith tracing
export LANGSMITH_API_KEY=your_langsmith_api_key
export LANGSMITH_TRACING=true
export LANGSMITH_PROJECT=your-project-name

CLI Configuration

# Set default values
imagen config set default_model gemini-2.0-flash-exp
imagen config set temperature 0.8
imagen config set aspect_ratio 16:9
imagen config set safety_settings relaxed

# View configuration
imagen config list

# Configuration location
imagen config path  # Shows: ~/.config/imagen/config.yaml

Configuration Precedence

Values are resolved in order (highest to lowest priority):

1. Command-line flags
2. Environment variables
3. Config file (~/.config/imagen/config.yaml)
4. Default values

Python API Reference

For complete API documentation with detailed examples, see LIBRARY.md.

Quick reference:

GeminiImageGenerator

generator = GeminiImageGenerator(
    model_name="gemini-2.5-flash-image",  # Image generation model (default)
    api_key=None,                         # Auto-loads from GOOGLE_API_KEY env var
    log_images=True                       # Enable LangSmith logging
)

generate() Method

result = await generator.generate(
    prompt: str,                           # Main prompt (required)
    system_prompt: Optional[str] = None,   # System instructions
    input_images: Optional[List] = None,   # Input images
    temperature: Optional[float] = None,   # Sampling temperature (0.0-1.0)
    aspect_ratio: Optional[str] = None,    # e.g., "16:9"
    safety_settings: Optional[List] = None,# Safety filtering
    output_images: Optional[List] = None,  # Generate images
    output_text: bool = False,             # Generate text
    metadata: Optional[Dict] = None,       # LangSmith metadata
    tags: Optional[List] = None            # LangSmith tags
) -> GenerationResult

See LIBRARY.md for full type definitions, parameter details, and usage examples.

Examples

See the examples/ directory for complete working examples:

• basic_generation.py - Simple text-to-image
• image_analysis.py - Analyze images
• labeled_inputs.py - Use labeled images
• s3_integration.py - S3 upload/download
• langsmith_tracing.py - Enable tracing

Documentation

• LIBRARY.md - Python library documentation, API reference, integration examples
• ADVANCED_USAGE.md - Advanced features, S3, LangSmith, scripting, automation
• docs/SAFETY_FILTERING.md - Safety filtering configuration and details
• CONTRIBUTING.md - Development setup, testing, contributing guidelines

Pricing

Image Generation (gemini-2.5-flash-image)

• Cost: $30/1M output tokens
• Per Image: ~$0.039 (1290 tokens at 1024x1024)

Text Model (gemini-2.5-flash)

• Input: $0.30/1M tokens
• Output: $1.20/1M tokens

Limitations

• Multiple images: Gemini may not always generate the exact number requested
• Structured output: Only available with text model (separate call required)
• Rate limits (free tier): 10 requests/minute, 1500/day

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

License

MIT License - see LICENSE for details.

Acknowledgments

• Built on google-genai - Google's unified GenAI SDK
• Uses langsmith for tracing
• S3 integration via boto3
• Type validation with pydantic v2
• CLI framework with click

Support

• Issues: GitHub Issues
• Documentation: This README and linked documentation files
• Examples: examples/ directory

---

Made with ❤️ by Aviad Rozenhek