ComfyUI

Acerca de

ComfyUI is an experimental server that seamlessly integrates ComfyUI workflows with Model Context Protocol (MCP) clients. It streamlines the management and execution of generative AI tasks by discovering workflow templates, cataloguing model assets like checkpoints and LoRAs, and enabling high-level parameter adjustments without direct JSON editing. The server facilitates workflow execution through the ComfyUI API, offering optional live streaming of prompt updates, making it ideal for integration with AI agents and IDEs.

Características Principales

• Perform high-level mutations to prompts, samplers, and resolution without editing graph JSON.
• Catalogue model assets including checkpoints, LoRAs, VAEs, and embeddings.
• Flexible configuration via TOML files and environment variables for base URL, directories, and feature toggles.
• Discover ComfyUI workflow templates and generate human-readable summaries.
• Execute workflows through the ComfyUI API with optional live streaming of updates.
• 0 GitHub stars

Casos de Uso

• Provide a programmatic interface for managing and running ComfyUI tasks from external applications.
• Automate the execution of generative AI prompts and asset management.
• Integrate ComfyUI workflows with MCP-aware IDEs and AI agents (e.g., Cursor, Claude Desktop).