Gpt Image MCP

🚀 Image Gen MCP Server

Empowering Universal Image Generation for AI Chatbots

Traditional AI chatbot interfaces are limited to text-only interactions, regardless of how powerful their underlying language models are. Image Gen MCP Server bridges this gap by enabling any LLM-powered chatbot client to generate professional-quality images through the standardized Model Context Protocol (MCP).

Whether you're using Claude Desktop, a custom ChatGPT interface, Llama-based applications, or any other LLM client that supports MCP, this server democratizes access to multiple AI image generation models including OpenAI's gpt-image-1, dall-e-3, dall-e-2, and Google's Imagen series (imagen-4, imagen-4-ultra, imagen-3), transforming text-only conversations into rich, visual experiences.

⚠️ Important Note

This project uses UV for fast, reliable Python package management. UV provides better dependency resolution, faster installs, and proper environment isolation compared to traditional pip/venv workflows.

🚀 Quick Start

Prerequisites

• Python 3.10+
• UV package manager
• OpenAI API key (for OpenAI models)
• Google Gemini API key (for Gemini models, optional)

Installation

1. Clone and setup:

   git clone <repository-url>
   cd image-gen-mcp
   uv sync
   

   💡 Usage Tip

   This project uses UV for fast, reliable Python package management. UV provides better dependency resolution and faster installs compared to pip.

2. Configure environment:

   cp .env.example .env
   # Edit .env and add your API keys:
   # - PROVIDERS__OPENAI__API_KEY for OpenAI models
   # - PROVIDERS__GEMINI__API_KEY for Gemini models (optional)
   

3. Test the setup:

   uv run python scripts/dev.py setup
   uv run python scripts/dev.py test
   

Running the Server

Development Mode

# HTTP transport for web development and testing
./run.sh dev

# HTTP transport with development tools (Redis Commander)
./run.sh dev --tools

# STDIO transport for Claude Desktop integration  
./run.sh stdio

# Production deployment with monitoring
./run.sh prod

Manual Execution

# STDIO transport (default) - for Claude Desktop
uv run python -m gpt_image_mcp.server

# HTTP transport - for web deployment
uv run python -m gpt_image_mcp.server --transport streamable-http --port 3001

# SSE transport - for real-time applications
uv run python -m gpt_image_mcp.server --transport sse --port 8080

# With custom configuration
uv run python -m gpt_image_mcp.server --config /path/to/.env --log-level DEBUG

# Enable CORS for web development
uv run python -m gpt_image_mcp.server --transport streamable-http --cors

Command Line Options

uv run python -m gpt_image_mcp.server --help

Image Gen MCP Server - Generate and edit images using OpenAI's gpt-image-1 model

options:
  --config PATH         Path to configuration file (.env format)
  --log-level LEVEL     Set logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
  --transport TYPE      Transport method (stdio, sse, streamable-http)
  --port PORT          Port for HTTP transports (default: 3001)
  --host HOST          Host address for HTTP transports (default: 127.0.0.1)
  --cors               Enable CORS for web deployments
  --version            Show version information
  --help               Show help message

Examples:
  # Claude Desktop integration
  uv run python -m gpt_image_mcp.server

  # Web deployment with Redis cache
  uv run python -m gpt_image_mcp.server --transport streamable-http --port 3001

  # Development with debug logging and tools
  uv run python -m gpt_image_mcp.server --log-level DEBUG --cors

MCP Client Integration

This server works with any MCP-compatible chatbot client. Here are configuration examples:

Claude Desktop (Anthropic)

{
  "mcpServers": {
    "image-gen-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/image-gen-mcp",
        "run",
        "image-gen-mcp"
      ],
      "env": {
        "OPENAI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Continue.dev (VS Code Extension)

{
  "mcpServers": {
    "gpt-image": {
      "command": "uv",
      "args": ["--directory", "/path/to/image-gen-mcp", "run", "image-gen-mcp"],
      "env": {
        "OPENAI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Custom MCP Clients

For other MCP-compatible applications, use the standard MCP STDIO transport:

uv run python -m gpt_image_mcp.server

💡 Usage Tip

This server follows the standard MCP protocol, ensuring compatibility with current and future MCP-enabled clients across the AI ecosystem.

✨ Features

• 🌐 Universal Compatibility: Works with any MCP-enabled LLM client
• 🔄 Seamless Integration: No context switching or workflow interruption
• ⚡ Standardized Protocol: One server, multiple client support
• 🎨 Multi-Provider Support: Access to OpenAI and Google's latest image generation models
• 🔧 Unified Interface: Single API for multiple AI providers with automatic model discovery

🎨 Multi-Provider Image Generation

• Multiple AI Models: Support for OpenAI (gpt-image-1, dall-e-3, dall-e-2) and Google Gemini (imagen-4, imagen-4-ultra, imagen-3)
• Text-to-Image: Generate high-quality images from text descriptions
• Image Editing: Edit existing images with text instructions (OpenAI models)
• Multiple Formats: Support for PNG, JPEG, and WebP output formats
• Quality Control: Auto, high, medium, and low quality settings
• Background Control: Transparent, opaque, or auto background options
• Dynamic Model Discovery: Query available models and capabilities at runtime

🔗 MCP Integration

• FastMCP Framework: Built with the latest MCP Python SDK
• Multiple Transports: STDIO, HTTP, and SSE transport support
• Structured Output: Validated tool responses with proper schemas
• Resource Access: MCP resources for image retrieval and management
• Prompt Templates: 10+ built-in templates for common use cases

💾 Storage & Caching

• Local Storage: Organized directory structure with metadata
• URL-based Access: Transport-aware URL generation for images
• Dual Access: Immediate base64 data + persistent resource URIs
• Smart Caching: Memory-based caching with TTL and Redis support
• Auto Cleanup: Configurable file retention policies

🚀 Production Deployment

• Docker Support: Production-ready Docker containers
• Multi-Transport: STDIO for Claude Desktop, HTTP for web deployment
• Reverse Proxy: Nginx configuration with rate limiting
• Monitoring: Grafana and Prometheus integration
• SSL/TLS: Automatic certificate management with Certbot

🛠️ Development Features

• Type Safety: Full type hints with Pydantic models
• Error Handling: Comprehensive error handling and logging
• Configuration: Environment-based configuration management
• Testing: Pytest-based test suite with async support
• Dev Tools: Hot reload, Redis Commander, debug logging

📦 Installation

The installation steps are included in the "Quick Start" section above.

💻 Usage Examples

Basic Usage

# Use via MCP client
result = await session.call_tool(
    "generate_image",
    arguments={
        "prompt": "A beautiful sunset over mountains, digital art style",
        "quality": "high",
        "size": "1536x1024",
        "style": "vivid"
    }
)

Advanced Usage

# Get optimized prompt for social media
prompt_result = await session.get_prompt(
    "social_media_prompt",
    arguments={
        "platform": "instagram",
        "content_type": "product announcement",
        "brand_style": "modern minimalist"
    }
)

# Access via resource URI
image_data = await session.read_resource("generated-images://img_20250630143022_abc123")

# Check recent images
history = await session.read_resource("image-history://recent?limit=5")

# Storage statistics
stats = await session.read_resource("storage-stats://overview")

📚 Documentation

Available Tools

list_available_models

List all available image generation models and their capabilities.

Returns: Dictionary with model information, capabilities, and provider details.

generate_image

Generate images from text descriptions using any supported model.

Parameters:

• prompt (required): Text description of desired image
• model (optional): Model to use (e.g., "gpt-image-1", "dall-e-3", "imagen-4")
• quality: "auto" | "high" | "medium" | "low" (default: "auto")
• size: "1024x1024" | "1536x1024" | "1024x1536" (default: "1536x1024")
• style: "vivid" | "natural" (default: "vivid")
• output_format: "png" | "jpeg" | "webp" (default: "png")
• background: "auto" | "transparent" | "opaque" (default: "auto")

Note: Parameter availability depends on the selected model. Use list_available_models to check capabilities.

edit_image

Edit existing images with text instructions.

Parameters:

• image_data (required): Base64 encoded image or data URL
• prompt (required): Edit instructions
• mask_data: Optional mask for targeted editing
• size, quality, output_format: Same as generate_image

Available Resources

• generated-images://{image_id} - Access specific generated images
• image-history://recent - Browse recent generation history
• storage-stats://overview - Storage usage and statistics
• model-info://gpt-image-1 - Model capabilities and pricing

Prompt Templates

Built-in templates for common use cases:

• Creative Image: Artistic image generation
• Product Photography: Commercial product images
• Social Media Graphics: Platform-optimized posts
• Blog Headers: Article header images
• OG Images: Social media preview images
• Hero Banners: Website hero sections
• Email Headers: Newsletter headers
• Video Thumbnails: YouTube/video thumbnails
• Infographics: Data visualization images
• Artistic Style: Specific art movement styles

Configuration

Configure via environment variables or .env file:

# =============================================================================
# Provider Configuration
# =============================================================================
# OpenAI Provider (default enabled)
PROVIDERS__OPENAI__API_KEY=sk-your-openai-api-key-here
PROVIDERS__OPENAI__BASE_URL=https://api.openai.com/v1
PROVIDERS__OPENAI__ORGANIZATION=org-your-org-id
PROVIDERS__OPENAI__TIMEOUT=300.0
PROVIDERS__OPENAI__MAX_RETRIES=3
PROVIDERS__OPENAI__ENABLED=true

# Gemini Provider (default disabled)
PROVIDERS__GEMINI__API_KEY=your-gemini-api-key-here
PROVIDERS__GEMINI__BASE_URL=https://generativelanguage.googleapis.com/v1beta/
PROVIDERS__GEMINI__TIMEOUT=300.0
PROVIDERS__GEMINI__MAX_RETRIES=3
PROVIDERS__GEMINI__ENABLED=false
PROVIDERS__GEMINI__DEFAULT_MODEL=imagen-4

# =============================================================================
# Image Generation Settings
# =============================================================================
IMAGES__DEFAULT_MODEL=gpt-image-1
IMAGES__DEFAULT_QUALITY=auto
IMAGES__DEFAULT_SIZE=1536x1024
IMAGES__DEFAULT_STYLE=vivid
IMAGES__DEFAULT_MODERATION=auto
IMAGES__DEFAULT_OUTPUT_FORMAT=png
# Base URL for image hosting (e.g., https://cdn.example.com for nginx/CDN)
IMAGES__BASE_HOST=

# =============================================================================
# Server Configuration
# =============================================================================
SERVER__NAME=Image Gen MCP Server
SERVER__VERSION=0.1.0
SERVER__PORT=3001
SERVER__HOST=127.0.0.1
SERVER__LOG_LEVEL=INFO
SERVER__RATE_LIMIT_RPM=50

# =============================================================================
# Storage Configuration
# =============================================================================
STORAGE__BASE_PATH=./storage
STORAGE__RETENTION_DAYS=30
STORAGE__MAX_SIZE_GB=10.0
STORAGE__CLEANUP_INTERVAL_HOURS=24

# =============================================================================
# Cache Configuration
# =============================================================================
CACHE__ENABLED=true
CACHE__TTL_HOURS=24
CACHE__BACKEND=memory
CACHE__MAX_SIZE_MB=500
# CACHE__REDIS_URL=redis://localhost:6379

🔧 Technical Details

The server follows a modular, production-ready architecture:

Core Components

• Server Layer (server.py): FastMCP-based MCP server with multi-transport support
• Configuration (config/): Environment-based settings management with validation
• Tool Layer (tools/): Image generation and editing capabilities
• Resource Layer (resources/): MCP resources for data access and model registry
• Storage Manager (storage/): Organized local image storage with cleanup
• Cache Manager (utils/cache.py): Memory and Redis-based caching system

Multi-Provider Architecture

• Provider Registry (providers/registry.py): Centralized provider and model management
• Provider Base (providers/base.py): Abstract base class for all providers
• OpenAI Provider (providers/openai.py): OpenAI API integration with retry logic
• Gemini Provider (providers/gemini.py): Google Gemini API integration
• Type System (types/): Pydantic models for type safety
• Validation (utils/validators.py): Input validation and sanitization

Infrastructure

• Prompt Templates (prompts/): Template system for optimized prompts
• Dynamic Model Discovery: Runtime model capability detection
• Parameter Translation: Automatic parameter mapping between providers

Deployment

• Docker Support: Development and production containers
• Multi-Transport: STDIO, HTTP, SSE transport layers
• Monitoring: Prometheus metrics and Grafana dashboards
• Reverse Proxy: Nginx configuration with SSL and rate limiting

📄 License

MIT License - see LICENSE file for details.

🖼️ Visual Showcase

Real-World Usage

Claude Desktop seamlessly generating images through MCP integration

Generated Examples

📋 Use Cases & Applications

🎯 Content Creation Workflows

• Bloggers & Writers: Generate custom illustrations directly in writing tools
• Social Media Managers: Create platform-specific graphics without leaving chat interfaces
• Marketing Teams: Rapid prototyping of visual concepts during brainstorming sessions
• Educators: Generate teaching materials and visual aids on-demand

🚀 Development & Design

• UI/UX Designers: Quick mockup generation during design discussions
• Frontend Developers: Placeholder and concept images within development environments
• Technical Writers: Custom diagrams and illustrations for documentation
• Product Managers: Visual concept communication in any LLM-powered tool

🏢 Enterprise Integration

• Customer Support: Generate visual explanations and guides
• Sales Teams: Custom presentation materials tailored to client needs
• Training Programs: Visual learning materials created in conversational interfaces
• Internal Tools: Add image generation to existing LLM-powered applications

🎨 Creative Industries

• Game Developers: Concept art and asset ideation
• Film & Media: Storyboard and concept visualization
• Architecture: Quick visual references and mood boards
• Advertising: Campaign concept development

💰 Cost Estimation

• Text Input: ~$5 per 1M tokens
• Image Output: ~$40 per 1M tokens (~1750 tokens per image)
• Typical Cost: ~$0.07 per image generation

🛡️ Error Handling

Comprehensive error handling includes:

• API rate limiting and retries
• Invalid parameter validation
• Storage error recovery
• Cache failure fallbacks
• Detailed error logging

🔐 Security

Security features include:

• OpenAI API key protection
• Input validation and sanitization
• File system access controls
• Rate limiting protection
• No credential exposure in logs

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests for new functionality
5. Run the test suite
6. Submit a pull request

🆘 Support

For issues and questions:

1. Check the troubleshooting guide
2. Review common issues
3. Open an issue on GitHub

🚢 Deployment

Production Deployment

The server supports production deployment with Docker, monitoring, and reverse proxy:

# Quick production deployment
./run.sh prod

# Manual Docker Compose deployment
docker-compose -f docker-compose.prod.yml up -d

Production Stack includes:

• Image Gen MCP Server: Main application container
• Redis: Caching and session storage
• Nginx: Reverse proxy with rate limiting (configured separately)
• Prometheus: Metrics collection
• Grafana: Monitoring dashboards

Access Points:

• Main Service: http://localhost:3001 (behind proxy)
• Grafana Dashboard: http://localhost:3000
• Prometheus: http://localhost:9090 (localhost only)

VPS Deployment

For VPS deployment with SSL, monitoring, and production hardening:

# Download deployment script
wget https://raw.githubusercontent.com/your-repo/image-gen-mcp/main/deploy/vps-setup.sh
chmod +x vps-setup.sh
./vps-setup.sh

Features included:

• Docker containerization
• Nginx reverse proxy with SSL
• Automatic certificate management (Certbot)
• System monitoring and logging
• Firewall configuration
• Automatic backups

See VPS Deployment Guide for detailed instructions.

Docker Configuration

Available Docker Compose profiles:

# Development with HTTP transport
docker-compose -f docker-compose.dev.yml up

# Development with Redis Commander
docker-compose -f docker-compose.dev.yml --profile tools up

# STDIO transport for desktop integration
docker-compose -f docker-compose.dev.yml --profile stdio up

# Production with monitoring
docker-compose -f docker-compose.prod.yml up -d

🔮 The Future of AI Integration

The Model Context Protocol represents a paradigm shift towards standardized AI tool integration. As more LLM clients adopt MCP support, servers like this one become increasingly valuable by providing universal capabilities across the entire ecosystem.

Current MCP Adoption

• ✅ Claude Desktop (Anthropic) - Full MCP support
• ✅ Continue.dev - VS Code extension with MCP integration
• ✅ Zed Editor - Built-in MCP support for coding workflows
• 🚀 Growing Ecosystem - New clients adopting MCP regularly

Vision

A future where AI capabilities are modular, interoperable, and user-controlled rather than locked to specific platforms.

🌟 Building the Universal AI Ecosystem

Democratizing advanced AI capabilities across all platforms through the power of the Model Context Protocol. One server, infinite possibilities.