ComfyUI-Copilot: AI Assistance for ComfyUI Workflows

Overview

ComfyUI-Copilot is an AI-powered assistant built on ComfyUI, the node-based interface for image-generation workflows. It aims to make workflow development less tedious by helping users create, debug, revise, and tune workflows from within ComfyUI.

The project is aimed at ComfyUI users who want conversational help while building image-generation workflows. Its README says the hosted API service has been suspended; Agent-related features remain available when users provide their own API key and base URL. The README also says node information queries, job recommendations, and workflow generation will no longer be available soon, so check the current project guidance before relying on those service-backed capabilities.

Key Features

• Generate workflow suggestions from a text description, including workflows from a library and AI-generated options, as described in the README.
• Analyze workflow errors, identify parameter or connection problems, and suggest fixes.
• Rewrite the current canvas based on instructions to change parameters, add nodes, or adjust workflow logic.
• Batch-run parameter combinations and compare generated results through the GenLab interface.
• Recommend nodes and downstream subgraphs based on a user's request or current canvas.
• Query selected nodes for descriptions, parameter information, and usage guidance.
• Find base models and LoRAs based on text requirements.

Use Cases

• ComfyUI beginners can ask for workflow ideas instead of assembling every node from scratch.
• Workflow developers can use the debugging assistant to investigate broken connections or incorrect parameters.
• Image creators can ask the copilot to modify an existing canvas when requirements change.
• Users comparing generation settings can batch-run parameter combinations and inspect the resulting images.
• ComfyUI users can look up node behavior or discover related nodes while working on a canvas.

Getting Started

Install the custom node in ComfyUI's custom_nodes directory and install its Python dependencies:

cd ComfyUI/custom_nodes
git clone https://github.com/AIDC-AI/ComfyUI-Copilot.git
cd ComfyUI-Copilot
pip install -r requirements.txt

Then start ComfyUI and activate Copilot from its interface. See the README for platform-specific installation steps and configuration details.

Considerations

• The hosted API service is suspended. Agent features require users to supply their own API key and base URL.
• The README warns that node information queries, job recommendations, and workflow generation will no longer be available soon. Their present status should be verified before adoption.
• The project is marked maintained and has had recent commits, but it has had no release in over a year. No issues were closed in the measured 90-day period, and the repository reports no automated tests.
• Installation requires Python dependencies in addition to ComfyUI. The provided requirements do not identify a separate database or Docker setup.
• The repository uses the MIT license.