Mcpserver

You can run a local MCP server that accepts your own facial images and ID images, blends them via a configured ComfyUI workflow, and saves the merged result locally. This enables automated, repeatable face-background fusion with your own images and keeps the output on your machine.

How to use

Start by installing dependencies, then run the MCP server and connect your MCP client (such as Claude) to it. You will provide image paths and a destination folder, and the system will perform face-background fusion using the Flux model and associated components.

How to install

Prerequisites: ensure Python is installed on your system.

# 1. Install dependencies
pip install -r requirements.txt

# 2. Run the MCP server
python server.py

Additional sections

Configuration notes: you can use the provided client configuration example to connect to your local MCP server. The client example shows how to point the client’s MCP connection to your server URL or address. You may adjust the affine transform target area and, if needed, the prompt parameters to influence the fusion result.

Example client connection snippet for the Flux MCP channel shows how to specify the MCP server address within the client configuration. Use your actual local server address when you configure the client.

Important notes: first run may take longer because large models are loaded on startup. You might encounter transient errors during model loading, but the final fusion result will typically save correctly.

Available tools

APersonMaskGenerator

A plugin used to generate person-specific masks for selective fusion within the ComfyUI workflow.

FluxFillModel

A model used to fill or harmonize the fused image background during the fusion process.