Comfy MCP Public Beta Lets AI Agents Automate ComfyUI Image and Video Generation
Comfy MCP removes the need for users to manually build complex ComfyUI node workflows, handing that task to an AI agent instead.
Reporting from 1 sources: GIGAZINE.
ComfyOrg has released the public beta of Comfy MCP, a system that lets AI agents like Claude Code and Codex operate ComfyUI on Comfy Cloud. Users add "@comfy" to agent instructions to build workflows and generate images or video. Support for the local ComfyUI version is on the roadmap, with volunteer MCP servers as a temporary workaround.
The public beta of Comfy MCP, announced Tuesday, connects AI agents such as Claude Code and Codex directly to Comfy Cloud. Once installed, a user adds "@comfy" to their instructions and the agent automatically builds the appropriate workflow and runs image or video generation. The system supports open models like Anima and LTX-2.3 as well as closed models like ChatGPT Images 2.0. Comfy MCP can also pull data from external apps like Notion for use in generation.
Currently Comfy MCP works only with Comfy Cloud, not the desktop version of ComfyUI. The development team has local support on the roadmap and is directing users to volunteer-built MCP servers in the meantime. Alongside the release, ComfyOrg published a Skills repository and a beta of comfy-cli, a command-line tool for running ComfyUI.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.