ComfyUI MCP Server
Enables Claude Desktop to interact with local ComfyUI installations for AI-powered image generation, including workflow management, model selection, real-time monitoring, and custom workflow execution through natural language.
README
ComfyUI MCP Server
A Model Context Protocol (MCP) server that enables Claude Desktop to interact with your local ComfyUI installation for AI-powered image generation.
Features
- 15 MCP Tools for complete ComfyUI control
- Template-based generation (Flux, SD1.5, SDXL, img2img)
- Custom workflow execution with smart parameter overrides
- Real-time progress monitoring via WebSocket
- Model management (list checkpoints, LoRAs, VAEs, etc.)
- Workflow library for saving and reusing workflows
- Queue management (status, cancel, clear)
- Image upload/retrieval with full Windows path support
Prerequisites
- Node.js v20 or higher
- ComfyUI installed and running at
http://127.0.0.1:8188 - Claude Desktop (for MCP integration)
- Windows 11 (as per specification)
Installation
1. Install Dependencies
[!NOTE] The file paths in the examples below must be replaced with the correct paths for your system.
cd [Path to your ComfyUI MCP Server]
npm install
2. Configure Paths
Edit config.json to match your ComfyUI installation:
{
"comfyui": {
"installation_path": "[Path to your ComfyUI portable installation]"
}
}
3. Build the Server
npm run build
4. Configure Claude Desktop
Edit %APPDATA%\Claude\claude_desktop_config.json:
{
"mcpServers": {
"comfyui": {
"command": "node",
"args": [
"[Path to your ComfyUI MCP Server]\\dist\\index.js"
],
"env": {
"COMFYUI_CONFIG": "[Path to your ComfyUI MCP Server]\\config.json"
}
}
}
}
5. Restart Claude Desktop
The ComfyUI tools will now be available in Claude Desktop.
Available Tools
Generation
comfy_submit_workflow- Submit custom workflow JSON with overridescomfy_generate_simple- Quick generation using templates
Status & Monitoring
comfy_get_status- Check generation status and outputscomfy_wait_for_completion- Wait for generation to complete
Model Management
comfy_list_models- List available models, LoRAs, VAEs
Workflow Library
comfy_save_workflow- Save workflow to librarycomfy_load_workflow- Load saved workflowcomfy_list_workflows- List all saved workflowscomfy_delete_workflow- Delete workflow from library
Queue Management
comfy_get_queue- Get current queue statuscomfy_cancel_generation- Cancel generationcomfy_clear_queue- Clear pending queue items
Utilities
comfy_upload_image- Upload image to ComfyUI input foldercomfy_get_output_images- List recent output images
Usage Examples
Simple Text-to-Image Generation
Ask Claude:
Generate an image of a sunset over mountains using Flux
Claude will use comfy_generate_simple with the flux_txt2img template.
Custom Workflow Execution
Use my chrono_edit workflow to animate this product image
Claude will:
- Load your workflow with
comfy_load_workflow - Upload the image with
comfy_upload_image - Submit with
comfy_submit_workflowand parameter overrides
Check Available Models
What Flux models do I have available?
Claude will use comfy_list_models with filter="flux".
Configuration
Template Defaults
Edit config.json to customize template defaults:
{
"templates": {
"flux_txt2img": {
"default_model": "flux_dev.safetensors",
"default_steps": 20,
"default_cfg": 3.5
}
}
}
Workflow Library Path
Workflows are saved to:
[Path to your ComfyUI portable installation]\ComfyUI\user\default\workflows\mcp_library
Troubleshooting
"Cannot connect to ComfyUI"
- Ensure ComfyUI is running:
run_nvidia_gpu.bat - Check ComfyUI is accessible at
http://127.0.0.1:8188 - Verify port 8188 is not blocked by firewall
"Model not found"
- Run
comfy_list_modelsto see available models - Check model file exists in the correct folder
- Verify model name spelling matches exactly
"Workflow validation failed"
- Test workflow in ComfyUI UI first
- Check all node connections are valid
- Ensure all required models are available
Permission Errors
- Check folder permissions on ComfyUI directories
- Run Claude Desktop as administrator if needed
- Verify paths in
config.jsonare accessible
Development
Build in Watch Mode
npm run dev
Clean Build
npm run clean
npm run build
Architecture
Claude Desktop (stdio) → MCP Server → ComfyUI API (HTTP/WebSocket)
↓
File System
- Models
- Input/Output
- Workflow Library
The MCP server runs as a separate Node.js process and communicates with ComfyUI via its HTTP API and WebSocket connections. It does not modify any ComfyUI files.
License
MIT
Support
For issues and questions, refer to the specification document or ComfyUI API documentation.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
E2B
Using MCP to run code via e2b.