🎨 opc-image-gen-mcp

MCP Server wrapping Stable Diffusion WebUI API (164K+ stars) — AI image generation for agents

Overview

opc-image-gen-mcp connects to a running Stable Diffusion WebUI (AUTOMATIC1111) instance via its REST API, enabling AI agents to generate images from text prompts, transform images with img2img, and list available models.

Tools

| Tool | Description | |------|-------------| | txt2img | Generate images from text prompts with full parameter control | | img2img | Transform existing images using prompts and denoising strength | | models_list | List all available Stable Diffusion model checkpoints |

Prerequisites

A running Stable Diffusion WebUI with --api flag:

# Clone and launch SD WebUI
git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui
cd stable-diffusion-webui
./webui.sh --api --listen

Set the API URL:

export SD_API_URL=http://localhost:7860

Installation

npm install opc-image-gen-mcp

Usage

As an MCP server

{
  "mcpServers": {
    "image-gen": {
      "command": "npx",
      "args": ["opc-image-gen-mcp"],
      "env": { "SD_API_URL": "http://localhost:7860" }
    }
  }
}

Example: txt2img

Input:  { "prompt": "a beautiful sunset over mountains, oil painting", "steps": 30, "width": 768, "height": 512 }
Output: Image saved: ./output/txt2img_1700000000.png

Example: img2img

Input:  { "input_image": "./photo.jpg", "prompt": "make it a watercolor painting", "denoising_strength": 0.6 }
Output: Image saved: ./output/img2img_1700000000.png

Example: models_list

Input:  {}
Output: [{ title: "v1-5-pruned-emaonly", model_name: "v1-5-pruned-emaonly", ... }]

Features

• Full SD parameter control — Steps, CFG scale, seed, resolution, sampler
• txt2img + img2img — Both generation modes with base64 image handling
• Model discovery — List all installed checkpoints automatically
• Configurable endpoint — Connect to any SD WebUI instance via SD_API_URL

License

MIT