Skip to content

v0.2.0

Choose a tag to compare

@JacobFV JacobFV released this 01 Jun 20:31
· 3 commits to main since this release

๐ŸŽจ imGPT v0.2.0 Release Notes

Release Date: May 26, 2025
Repository: https://github.com/JacobFV/imgpt

๐Ÿš€ Major New Features

๐Ÿค– Multi-Model Support

We've added support for all OpenAI image generation models:

  • gpt-image-1 (default) - OpenAI's latest and most advanced image model
  • dall-e-3 - High-quality artistic images with style controls
  • dall-e-2 - Fast and cost-effective image generation
# Use different models
imgpt "A sunset" --model dall-e-3
imgpt "Quick sketch" --model dall-e-2

๐Ÿ“ Multiple Image Dimensions

Generate images in various aspect ratios and sizes:

gpt-image-1:

  • Square: 1024x1024
  • Landscape: 1536x1024
  • Portrait: 1024x1536

dall-e-3:

  • Square: 1024x1024
  • Landscape: 1792x1024
  • Portrait: 1024x1792

dall-e-2:

  • Small: 256x256
  • Medium: 512x512
  • Large: 1024x1024
# Generate in different orientations
imgpt "Landscape photo" --size 1792x1024
imgpt "Portrait painting" --size 1024x1536
imgpt "Thumbnail" --model dall-e-2 --size 256x256

๐ŸŽจ Quality & Style Controls

Fine-tune your image generation:

  • Quality levels: auto, high, medium, low, hd, standard
  • Style options: vivid, natural (DALL-E 3 only)
  • Output formats: PNG, JPEG, WebP (gpt-image-1 only)
# High-quality artistic image
imgpt "Abstract art" --model dall-e-3 --quality hd --style vivid

# Web-optimized JPEG
imgpt "Website banner" --format jpeg --quality medium

๐Ÿ›ก๏ธ Enhanced Validation & Intelligence

Smart Model Validation

  • Automatic validation of size compatibility with each model
  • Model-specific prompt length limits (32K for gpt-image-1, 4K for dall-e-3, 1K for dall-e-2)
  • Intelligent default sizes based on selected model

Better Error Messages

$ imgpt "test" --model dall-e-2 --size 2048x2048
โŒ Error: Size '2048x2048' is not valid for model 'dall-e-2'
Valid sizes for dall-e-2: 256x256, 512x512, 1024x1024

๐Ÿ“– Comprehensive Documentation Updates

Updated CLI Help

$ imgpt --help
Options:
  --model {gpt-image-1,dall-e-2,dall-e-3}    Model to use (default: gpt-image-1)
  --size SIZE                                 Image dimensions (e.g., 1024x1024)
  --quality {auto,high,medium,low,hd,standard} Image quality (default: high)
  --style {vivid,natural}                     DALL-E 3 style options
  --format {png,jpeg,webp}                    Output format (default: png)

Enhanced README

  • Detailed model comparison table
  • Comprehensive usage examples
  • Best practices for each model
  • Cost optimization tips

๐Ÿ”ง Technical Improvements

API Integration

  • Full support for OpenAI's latest image generation API parameters
  • Model-specific parameter handling
  • Proper response format handling for each model

Code Quality

  • Enhanced type hints and documentation
  • Improved error handling and validation
  • Extended test coverage for new features
  • Backward compatibility maintained

๐Ÿ“‹ Usage Examples

Single Image Generation

# Default high-quality generation
imgpt "A majestic eagle in flight"

# Artistic style with DALL-E 3
imgpt "Surreal landscape" --model dall-e-3 --style vivid --quality hd

# Fast generation for prototyping
imgpt "Logo concept" --model dall-e-2 --size 512x512

# Custom format and quality
imgpt "Web image" --format jpeg --quality medium --size 1536x1024

Batch Processing

# High-quality art generation
imgpt --dir ./art_prompts --model dall-e-3 --quality hd --style natural

# Thumbnail generation
imgpt --dir ./thumbnails --model dall-e-2 --size 256x256

# Web-optimized batch processing
imgpt --dir ./web_images --format webp --quality medium

๐Ÿ”„ Breaking Changes

None - This release maintains full backward compatibility. All existing commands will continue to work exactly as before.

๐Ÿ› Bug Fixes

  • Fixed output format handling for different models
  • Improved prompt length validation
  • Enhanced file extension handling for custom output formats

๐Ÿš€ Performance Improvements

  • Optimized API parameter building
  • Reduced unnecessary API calls through better validation
  • Improved error handling reduces retry overhead

๐Ÿ“ฆ Installation & Upgrade

New Installation

# Install with pipx (recommended)
pipx install imgpt

# Install with pip
pip install imgpt

Upgrade from v0.1.0

# Upgrade with pipx
pipx upgrade imgpt

# Upgrade with pip
pip install --upgrade imgpt

๐Ÿ™ Acknowledgments

  • OpenAI for their comprehensive image generation API
  • Community feedback that drove these feature requests
  • Contributors who helped test the new functionality

๐Ÿ”ฎ What's Next?

  • Batch size optimization for faster processing
  • Image editing and variation capabilities
  • Integration with popular design tools
  • Advanced prompt templating system

Full Changelog: v0.1.0...v0.2.0
Download: https://github.com/JacobFV/imgpt/releases/tag/v0.2.0

Made with โค๏ธ by Jacob Valdez


๐ŸŽจ imGPT v1.0.0 Release Notes

Release Date: May 26, 2025
Repository: https://github.com/JacobFV/imgpt

๐Ÿš€ Features

  • Generate images from text prompts using OpenAI's image-gpt model
  • Supports direct prompt input and batch processing of prompt files
  • Output images in PNG format
  • Simple CLI interface

๐Ÿ› Bug Fixes

  • Initial release, no known bugs

๐Ÿ“ฆ Installation

pip install imgpt