Getting started
Install ComfyArt into ComfyUI, open the canvas and make your first image.
Models
Which model files each feature needs, and where to put them.
Explore
The canvas
Pages, frames, layers, text and shapes, comments, history and export to PNG, PDF and PPTX.
Generate
Image and video generators with references, color palettes, base images and storyboards.
Image tools
Quick edit, upscale, split layers, remove background, eraser, edit text, multi-angle, inpaint and more.
Video
Animate an image, generate and extend clips, trim and edit, upscale to 4K.
Assets
My Assets, brand kits, characters, products and the reference board.
Agent
Chat with a design agent (Claude Code, Codex, Qwen Code, DeepSeek Harness or a local LLM) that works on your canvas.
Why ComfyArt
Fully local
Every model runs in your ComfyUI. Nothing is uploaded, and ComfyArt never stores credentials.
Open workflows
Each canvas action is a plain ComfyUI API workflow in
workflows/, built from core nodes wherever possible.Agent-ready
An MCP server lets the built-in agent, or any local MCP client, look at the canvas and edit it.