ComfyArt is a free, local Lovart-style design canvas with an AI agent, running on ComfyUI
ComfyArt is a free, open-source infinite design canvas that installs into ComfyUI and generates images, video, music, voiceovers and 3D models on your own GPU. A design agent such as Claude Code, Codex or a local LLM can work on the canvas through MCP, with no account or credits.
What it does
The canvas has pages, artboards, rich text, a pen tool, booleans, auto layout, layers and a brand kit, and exports PNG, JPG, PDF and editable PPTX. Images come from Qwen Image 2.1, Z-Image Turbo, FLUX.2 Klein, FLUX 3 Klein, MOSS-Image, Qwen Image 2512 or Ideogram 4, with references, depth, pose and line-art modes; video uses MiniMax H3 from text, first and last frames, up to 9 reference images or a storyboard, at 5 to 15 seconds, with motion control and lip sync. There are image tools (inpaint, expand, remove background, multi-angle, split layers, vectorize), Edit Elements to turn a flat poster back into editable layers, product mockups, music (MiniMax Music 3 or ACE-Step 1.5), sound effects, voiceover with voice cloning, image-to-3D, and a multi-clip video editor. Every action runs as a plain ComfyUI workflow, and jobs can go to other ComfyUI machines on your network.
Worth knowing
Install by cloning it into ComfyUI/custom_nodes and running pip install -r requirements.txt, then click the ComfyArt button in ComfyUI's top bar. The agent is off by default and uses a CLI you already have (Claude Code, Codex, Qwen Code, DeepSeek Harness) or any OpenAI-compatible local LLM with tool calling. Without a local GPU it can send jobs to Comfy Cloud, which needs a paid Comfy plan API key. ComfyArt is MIT, but the bundled tldraw canvas has its own licence and shows a small licence badge when served from localhost. Version 1.1.0 (27 October) added the top-bar button.