fm-mcp-comfyui-bridge

fm-mcp-comfyui-bridge

1
entertainment_and_mediaai_chatbot

The fm-mcp-comfyui-bridge is an MCP server implementation designed to interface with ComfyUI for image generation. It offers features such as image captioning and tag analysis, and is easily set up using the `uv` package manager. It provides API endpoints for seamless integration with other tools.

fm-mcp-comfyui-bridge

Features

  • Image generation using ComfyUI
  • Automatic caption generation for images
  • Tag analysis for generated images
  • Easy setup and launch with uv
  • Provides API endpoints as an MCP server

Requirements

  • Python 3.13+
  • Locally running ComfyUI
  • uv package manager
  • Locally running ollama with vision model

Installation

Ensure you have an environment with uv installed.

Usage

  • Configure as an MCP server in the agent tool settings.
  • Adjust ComfyUI's endpoints as needed.
  • Lora model settings need to be created for image generation.

Available Tools

  • generate_picture - Generates an image based on a prompt
  • get_picture - Retrieves PNG binary data of a specified image
  • get_caption - Retrieves the caption of a specified image
  • get_tag - Analyzes and retrieves WD1.4 tags from an image

Dependencies

  • MCP server framework
  • Requests, Hugging Face, Numpy, Ollama, ONNX Runtime, Pandas

License

This project is licensed under MIT. See for details.

Author