Skip to main content
Glama
krystian-ai

AI Image-Gen MCP Server

by krystian-ai
OverviewSchemaSourceRelated ServersScoreDiscussionsAdmin
Python
Remote

AI Image‑Gen MCP Server

Version 0.1.0 – MVP public release
Conforms to the Model Context Protocol spec (2025‑06‑18).

Python 3.11+ License: MIT Code style: black MCP

What's this?

A production‑ready MCP server that transforms state‑of‑the‑art image generators into plug‑and‑play tools for any MCP‑aware client. Currently shipping with DALL·E 3, DALL·E 2, and experimental GPT‑Image‑1 – all accessible through a unified interface.

Why MCP?

MCP is the USB‑C of AI context: one protocol, endless integrations. Ship one server, hook it into Claude Desktop, Claude Code, VS Code, or your own chatbot – the host handles UI, auth, and conversation flow.

Quick examples

You ask

The server delivers

"Design a cyberpunk logo for my startup"

High‑res PNG via DALL·E 3 with style presets

"Generate 5 variations of this product shot"

Batch generation via DALL·E 2 (n=5 support)

"Create concept art for a steampunk airship"

Artistic rendering with metadata and prompt history

If you can describe it, we can render it. 💫

Related MCP server: MCP OpenAI Image Generation Server

Core MCP Concepts

This server implements all three MCP primitives:

  1. Tools – generate_image with model selection, size, and style options

  2. Resources – Available models and their capabilities exposed as MCP resources

  3. Prompts – Built‑in templates for product_mockup and concept_art workflows

Feature Highlights

  • Multi‑Model Support – DALL·E 3 (default), DALL·E 2, and GPT‑Image‑1 via unified API

  • Smart Storage – Local cache with timestamped filenames and JSON metadata

  • Flexible Sizing – From 256×256 thumbnails to 1792×1024 widescreen masterpieces

  • Style Control – vivid or natural rendering (DALL·E 3)

  • Batch Generation – Create up to 10 variations per prompt (DALL·E 2)

  • Claude Integration – First‑class support for Desktop and Code editions

Architecture

graph TD
    Client["MCP Client (Claude Desktop/Code)"] -- JSON‑RPC 2.0 --> Server["Image‑Gen MCP Server"]
    Server --> Router["Model Router"]
    Router -->|OpenAI API| DALLE3["DALL·E 3"]
    Router -->|OpenAI API| DALLE2["DALL·E 2"]
    Router -->|Responses API| GPT["GPT‑Image‑1"]
    Server --> Storage["Local Storage + Metadata"]
    Storage --> Client

Quickstart

Prerequisites

  • Python 3.11+

  • OpenAI API key

  • Claude Desktop or Claude Code (for MCP integration)

Installation

git clone https://github.com/krystian-ai/ai-image-gen-mcp.git
cd ai-image-gen-mcp
python3.11 -m venv .venv && source .venv/bin/activate
pip install -e ".[image,dev]"

Configuration

cp .env.example .env
# Edit .env and add your OpenAI API key

Key settings:

OPENAI_API_KEY=sk-...
MODEL_DEFAULT=dall-e-3
CACHE_DIR=/tmp/ai-image-gen-cache

Run Standalone

# Via MCP CLI
mcp-imageserve stdio

# Direct execution
python -m ai_image_gen_mcp.server --transport=stdio

Claude Desktop Integration

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "ai-image-gen": {
      "command": "/path/to/ai-image-gen-mcp/.venv/bin/python",
      "args": ["-m", "ai_image_gen_mcp.server", "stdio"],
      "transport": "STDIO",
      "env": {
        "PYTHONPATH": "/path/to/ai-image-gen-mcp/src",
        "OPENAI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Claude Code Integration

Option 1: Project‑specific .mcp.json (Recommended)

Drop this in your project root:

{
  "ai-image-gen": {
    "command": "python",
    "args": ["-m", "ai_image_gen_mcp.server", "stdio"],
    "transport": "STDIO",
    "env": {
      "PYTHONPATH": "src",
      "OPENAI_API_KEY": "your-api-key-here"
    }
  }
}

Claude Code auto‑detects and loads it. ✨

Option 2: Global Config

Add to ~/.config/claude-code/settings.json for system‑wide access.

Model Capabilities

Model

Sizes

Styles

Batch (n)

Speed

Notes

DALL·E 3

1024×1024, 1792×1024, 1024×1792

vivid, natural

1

Fast

Best quality, default model

DALL·E 2

256×256, 512×512, 1024×1024

N/A

1-10

Fast

Good for variations

GPT‑Image‑1

Fixed (model‑determined)

N/A

1

Slow (20s+)

Experimental, may timeout

Usage Examples

Basic Generation

Generate a minimalist logo for a productivity app

With Parameters

Create a vivid 1792x1024 banner of a futuristic cityscape using dall-e-3

Batch Creation

Generate 5 variations of a coffee cup product photo using dall-e-2

Interactive HTML Demo

Explore the server's capabilities through our interactive web interface:

cd examples/html
open index.html  # macOS
# or
xdg-open index.html  # Linux
# or just open in your browser

The demo showcases:

  • Live Examples – Generated images with their prompts

  • Model Comparison – See outputs from DALL·E 3, DALL·E 2, and GPT-Image-1

  • Interactive Gallery – Carousel of stunning AI-generated artwork

  • Integration Guide – How to connect with Claude Desktop/Code

Perfect for visualizing what's possible before diving into the API!

Development

Testing

pytest                           # Full suite
pytest --cov=ai_image_gen_mcp   # Coverage report
python test_dalle.py            # Live API test

Code Quality

black src/        # Format
ruff check src/   # Lint
mypy src/         # Type check

Project Structure

ai-image-gen-mcp/
├── src/ai_image_gen_mcp/
│   ├── server.py          # FastMCP server entry
│   ├── models/            # Model implementations
│   └── config.py          # Environment config
├── tests/                 # Comprehensive test suite
├── examples/
│   └── html/              # Interactive web demo
├── assets/                # Logo images
└── .mcp.json             # Claude Code config

Troubleshooting

Issue

Solution

MCP not detected

Ensure .mcp.json exists in project root

API key errors

Check OPENAI_API_KEY in .env or environment

Import errors

Verify PYTHONPATH includes src/ directory

GPT‑Image‑1 timeouts

Known issue – use DALL·E models for reliability

Claude Desktop issues

Use full paths to venv Python executable

Roadmap

Version

Focus

Status

0.1

MVP with 3 models, local storage

✅ Shipped

0.2

S3/GCS storage, signed URLs

🚧 Planning

0.3

Stable Diffusion, ComfyUI integration

📋 Backlog

0.4

Inpainting, upscaling, style transfer

💭 Ideas

Contributing

Fork → feature branch → PR. Run pre-commit hooks. Keep the vibe technical but approachable.

License

MIT – see LICENSE.

Links

Available Tools

1 tool
generate_imageA

Generate images from text descriptions using AI models.

Args: prompt: Text description of the desired image style: Style preset (default, photorealistic, illustration) size: Image dimensions (1024x1024, 1792x1024, 1024x1792) n: Number of images to generate (currently only 1 supported) model: Specific model to use (dalle-3, dalle-2, gpt-image-1)

Returns: ImageGenerationResponse with image URLs and metadata

ParametersJSON Schema
NameRequiredDescriptionDefault
promptYes
styleNodefault
sizeNo1024x1024
nNo
modelNo

Output Schema

ParametersJSON Schema
NameRequiredDescription
image_urlsYesURLs or paths to generated images
promptYesThe prompt used for generation
modelYesModel used for generation
created_atYesISO 8601 timestamp of generation
messageNoUser-friendly message about the result

TDQS

A4/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

The description discloses that only 1 image is currently supported for the 'n' parameter, which is a helpful behavioral trait. However, it omits other important details such as rate limits, content filtering, or authentication requirements. Since no annotations are provided, the description should carry more burden.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise and well-structured with Args and Returns sections. It avoids unnecessary wording, though it could be slightly more front-loaded with the main purpose before the parameter list.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the presence of an output schema, the description adequately outlines the return value. It covers all parameters and the core functionality. Minor gaps in usage guidelines and behavioral traits prevent a perfect score.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 0% schema description coverage, the description adds complete meaning for all parameters by providing brief explanations (e.g., 'Text description of the desired image', 'Style preset (default, photorealistic, illustration)'). This fully compensates for the missing schema descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool generates images from text descriptions using AI models, which is a specific verb-resource combination. There are no sibling tools, so differentiation is not required.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No explicit guidance on when to use or not use this tool versus alternatives. The description implies general image generation but lacks exclusions or context for when other tools might be more appropriate.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

  1. 1 tool updatev0.1.0
    • First observedgenerate_image
Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A3.9/5.0
Disambiguation5/5

With only one tool, there is no risk of ambiguity. The tool has a clear and distinct purpose.

Naming Consistency5/5

The single tool name 'generate_image' follows a consistent verb_noun pattern, which is clear and predictable.

Tool Count3/5

A single tool for an image generation server is borderline. While it covers the core function, the surface feels thin compared to typical MCP servers with 3-15 tools.

Completeness2/5

The server only provides a create operation (generate_image). Missing get, update, delete, or list capabilities for generated images, which are significant gaps for managing outputs.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/krystian-ai/ai-image-gen-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server