ComfyUI-Copilot is an open source AI assistant plugin for the ComfyUI framework that lets you build, debug, and deploy AI image generation workflows using natural language commands instead of manually wiring node graphs.
The Problem
ComfyUI is a powerful tool for building AI image generation pipelines, but its node-based graph editor has a steep learning curve. Assembling a working workflow requires knowing which nodes exist, how they connect, and what parameters do what. Debugging a broken pipeline means tracing each connection manually. This creates a barrier for anyone who understands what they want to generate but not how to wire it up.
How ComfyUI-Copilot Solves It
ComfyUI-Copilot adds a natural language interface on top of ComfyUI's existing node system. Describe the pipeline you want, and it assembles the nodes. Ask it to fix an error, and it diagnoses the broken connection and suggests a correction. It integrates as a standard custom node, so it works alongside any existing ComfyUI workflow. Apache-2.0 license applies.
Key Features
- Natural language workflow construction: describe your AI pipeline and the plugin wires the nodes
- Error diagnosis and fix suggestions for broken node connections and parameter mismatches
- Model querying to find and load base models and LoRA adapters on demand
- Automated parameter tuning using ML optimization for critical workflow parameters
- Runs locally within ComfyUI: no data sent to external servers
Who It's For
ComfyUI-Copilot is best for AI image generation practitioners, researchers, and hobbyists who already use ComfyUI and want to build and debug complex workflows faster without manually configuring every node.
Compared to Manual Node Configuration
Unlike building ComfyUI workflows by hand, ComfyUI-Copilot translates natural language intent into node graphs automatically, reducing setup time for complex pipelines. Commercial AI coding assistants like GitHub Copilot do not integrate with ComfyUI's visual graph environment.
