Release v0.3.0

🍌 What's New

Model Upgrade

• Upgraded to Gemini 2.5 Flash Image - Now using the latest gemini-2.5-flash-image 🍌 model for improved image generation quality and performance

New Features

• Aspect Ratio Support - Added aspectRatio parameter to generate_image tool with 10 preset options:
  • Square: 1:1
  • Portrait: 2:3, 3:4, 4:5, 9:16
  • Landscape: 3:2, 4:3, 5:4, 16:9, 21:9

🔧 Technical Improvements

API Layer

• Extended GenerateImageParams with aspectRatio field
• Added imageConfig support for generateContent API
• Comprehensive type definitions for aspect ratio options

Validation

• Added aspect ratio validation logic
• Schema validation for MCP tool parameters
• Comprehensive test coverage for all aspect ratio values

📦 Installation

npm install mcp-image@0.3.0

---

Full Diff: v0.2.3...v0.3.0

Release v0.2.3

🔧 Improvements

• Enhanced Gemini API error handling with detailed safety filter detection
• Better error context for content filtering scenarios

🐛 Bug Fixes

• Fixed error message extraction from Gemini API text responses
• Resolved TypeScript strict mode compatibility issues

📦 Dependencies

• No dependency updates in this release

Release v0.2.2

Bug Fixes

• improve inputImagePath parameter description: Enhanced parameter description for better LLM understanding when using image-to-image generation
• increase maxTokens limit: Improved prompt generation by increasing token limits for better structured prompts

Changes

• Updated package version
• Improved parameter documentation for inputImagePath in MCP server

Release v0.2.1

Bug Fixes

Improved Prompt Generation

• Fixed: Increased maxTokens limit from 500 to 2000 in structured prompt generation
• Impact: Resolves issues with truncated prompts for complex image generation requests
• Details: The Gemini API's structured prompt generation now has sufficient token capacity to handle detailed, multi-layered prompts without truncation

Technical Details

• Modified src/business/structuredPromptGenerator.ts
• Adjusted Gemini 2.0 Flash API configuration parameters
• No breaking changes to existing API or functionality

Compatibility

• Fully backward compatible with v0.2.0
• No changes required for existing integrations

Release v0.2.0

🎯 What's New

Intelligent Prompt Enhancement

• Integrated Gemini 2.0 Flash for automatic prompt optimization before image generation
• Transforms simple prompts into rich, detailed descriptions with professional photography and artistic terminology
• Significantly improves image quality while maintaining user intent

Enhanced Image Editing

• Context-aware editing that preserves the original image's style and atmosphere
• Better consistency between source and edited images

🔧 Configuration

New Environment Variable

• SKIP_PROMPT_ENHANCEMENT=true - Bypass automatic prompt optimization for direct control over exact prompt wording

🚀 Improvements

• Two-stage processing pipeline: prompt optimization → image generation
• Better handling of multi-modal inputs (text + images)
• Optimized for both simple natural language and complex technical prompts
• Minimal additional latency for substantial quality improvements

📦 Technical Details

• Uses Gemini 2.0 Flash for prompt enhancement (minimal token usage)
• Continues to use Gemini 2.5 Flash Image Preview for image generation
• Fully backward compatible with v0.1.x configurations

---

Full Changelog: v0.1.1...v0.2.0

Release v0.1.1

🐛 Bug Fixes

• Exclude vitest.config.mjs from npm package to reduce package size

📦 Package

• Removed unnecessary test configuration files from published package
• Optimized package contents for production use

🔧 What's Changed

• Added vitest.config.mjs to .npmignore

Full Changelog: v0.1.0...v0.1.1

Release v0.1.0

We're excited to announce the initial release of MCP Image Generator - a powerful Model Context Protocol (MCP) server that brings AI image generation capabilities directly to Claude Code, Cursor, and other MCP-compatible AI tools.

✨ What's New

Core Features

• AI-Powered Image Generation: Generate stunning images from text prompts using Google's Gemini 2.5 Flash Image API
• Image Editing: Transform and modify existing images with natural language instructions
• Advanced Generation Options:
  • Multi-image blending for composite scenes
  • Character consistency across multiple generations
  • World knowledge integration for historically accurate content
• Multiple Output Formats: Support for PNG, JPEG, and WebP formats
• File-Based Output: Images are automatically saved as files for easy access and integration

Easy Integration

• One-Command Setup: Install and configure with a single Claude CLI command
• Multiple Platform Support: Works seamlessly with Claude Code and Cursor
• Flexible Configuration: Support for custom output directories and environment variables

Developer Experience

• TypeScript: Built with full TypeScript support for better development experience
• Comprehensive Testing: 100% test coverage with integration tests
• Quality Assurance: Automated linting, formatting, and code quality checks
• MCP Best Practices: Follows Model Context Protocol standards for optimal compatibility

🚀 Quick Start

claude mcp add mcp-image --env GEMINI_API_KEY=your-api-key --env IMAGE_OUTPUT_DIR=/path/to/images -- npx -y https://github.com/shinpr/mcp-image

📋 Requirements

• Node.js 20 or higher
• Gemini API key from Google AI Studio
• Claude Code, Cursor, or any MCP-compatible AI tool

🔧 Usage

Once installed, simply ask your AI assistant to generate images:

"Generate a serene mountain landscape at sunset with a lake reflection"
"Edit this photo to make the person face right"
"Create a medieval knight portrait with character consistency enabled"

🛠️ Technical Details

• Runtime: Node.js 20+
• Language: TypeScript 5.0+
• API: Google Gemini 2.5 Flash Image Preview
• Protocol: Model Context Protocol (MCP) 1.0
• Testing: Vitest with comprehensive coverage
• Code Quality: Biome for linting and formatting

📦 Installation Methods

Claude Code (Recommended)

claude mcp add mcp-image --env GEMINI_API_KEY=your-key -- npx -y https://github.com/shinpr/mcp-image

Manual Configuration

Add to your MCP configuration file with the server settings provided in our documentation.