Ideogram MCP Server

This server provides image generation capabilities using the Ideogram API via the Model Context Protocol (MCP), integrating seamlessly with tools like Claude Desktop and other MCP clients.

Key features:

Generate images based on text prompts (must be in English)
Customize with various models (V_1, V_1_TURBO, V_2, V_2_TURBO)
Control aspect ratio (1:1, 4:3, 16:9, etc.)
Add negative prompts to exclude unwanted elements
Generate multiple images at once (1-8)
Set image resolution from multiple options
Use seed values for reproducible results
Apply magic prompt feature (AUTO, ON, OFF)
Select rendering speed (TURBO, DEFAULT, QUALITY)
Apply specific styles using style codes or reference images
Choose style type (AUTO, GENERAL, REALISTIC, DESIGN)
Save generated images to specified directories
Apply blur masks to image edges

📦 Project Overview

A TypeScript tool that allows you to use Ideogram API (v3.0) via MCP server
Multi-function including image generation, style reference, magic prompt, aspect ratio, model selection, etc.
Can be used immediately with Claude Desktop and other MCP clients

Related MCP server: Image Generation MCP Server

⚡️ Quick Start

If you want to connect to Claude Desktop or other MCP clients at lightning speed,
Just copy and paste the JSON snippet below into your configuration file!

{ "mcpServers": { "ideogram": { "command": "npx", "args": [ "@sunwood-ai-labs/ideagram-mcp-server" ], "env": { "IDEOGRAM_API_KEY": "your_api_key_here" } } } }

🛠️ MCP Tool Specifications

generate_image

Parameter list (latest version)

Parameters	Type	explanation	Required/Optional	remarks
prompt	string	Image generation prompt (English recommended)	Required
aspect_ratio	string	Aspect ratio (e.g. "1x1", "16x9", "4x3", etc.)	any	15 types
resolution	string	Resolution (see official documentation, 69 types in total)	any
seed	integer	Random number seed (to ensure reproducibility)	any	0 to 2147483647
magic_prompt	string	Magic prompt ("AUTO"	"ON"	"OFF"
rendering_speed	string	Rendering speed for v3 ("TURBO"	"DEFAULT"	"QUALITY"
style_codes	string[]	8-character style code sequence	any
style_type	string	Style type ("AUTO"	"GENERAL"	"REALISTIC"
negative_prompt	string	Exclusions (English recommended)	any
num_images	number	Number of generated images (1 to 8)	any
style_reference	object	Style Reference (New in Ideogram 3.0)	any	Details below
└ urls	string[]	Reference image URL array (up to 3)	any
└ style_code	string	Style Code	any
└ random_style	boolean	Use random style	any
output_dir	string	Image storage directory (default: "docs")	any
base_filename	string	Base for saved filename (default: "ideogram-image")	any	Timestamp and ID assignment
blur_mask	boolean	Blur the edges of the image (set to true for mask compositing)	any	Default: false

📝 Usage example

const result = await use_mcp_tool({ server_name: "ideagram-mcp-server", tool_name: "generate_image", arguments: { prompt: "A beautiful sunset over mountains", aspect_ratio: "16x9", rendering_speed: "QUALITY", num_images: 2, style_reference: { urls: [ "https://example.com/ref1.jpg", "https://example.com/ref2.jpg" ], random_style: false }, blur_mask: true } });

🧑‍💻 Develop, build, and test

npm run build ... TypeScript build
npm run watch ... development mode (automatic build)
npm run lint ... Code linting
npm test ... run tests

🗂️ Directory structure

ideagram-mcp-server/ ├── assets/ ├── docs/ │ └── ideogram-image_2025-05-18T06-31-45-777Z.png ├── src/ │ ├── tools/ │ ├── types/ │ ├── utils/ │ ├── ideogram-client.ts │ ├── index.ts │ ├── server.ts │ └── test.ts ├── .env.example ├── package.json ├── tsconfig.json ├── README.md └── ...（省略）

📝 Contributions

Fork this repository
Create a new branch ( git checkout -b feature/awesome )
Commit changes (Commit messages should be in Japanese and emojis are recommended!)
Push and pull request creation

🚀 Deploy & Release

Automatic npm publishing with GitHub Actions
Version update → Automatic deployment by pushing tags

npm version patch|minor|major git push --follow-tags

For details, see docs/npm-deploy.md !

📄 License

MIT