Sign in

Nodes Browser

ComfyDeploy: How ComfyUI-Autogen works in ComfyUI?

What is ComfyUI-Autogen?

A collection of nodes for using Autogen with ComfyUI [a/AutoGen](https://github.com/microsoft/AutoGen): assistant agents, group chats, code executor, etc.

How to install it in ComfyDeploy?

Head over to the machine page

  1. Click on the "Create a new machine" button
  2. Select the Edit build steps
  3. Add a new step -> Custom Node
  4. Search for ComfyUI-Autogen and select it
  5. Close the build step dialig and then click on the "Save" button to rebuild the machine

ComfyUI-Autogen

A collection of nodes for using Autogen with ComfyUI

[AutoGen Stable]: assistant agents, group chats, code executor, etc.

Features

  1. Autogen Model

    • Initializes an OpenAI-like model via autogen_ext.models.
    • Lets you configure model name, base URL, API key, and optional capabilities like vision, function calling, and JSON output.

  2. Autogen Assistant Agent

    • Wraps an AssistantAgent from autogen_agentchat.agents.
    • Lets you specify tools (by scanning a tools folder), temperature, max tokens, etc.
    • Great for orchestrating more conversational or tool-using LLM tasks within ComfyUI.

  3. Autogen Code Executor

    • A node wrapping the CodeExecutorAgent from autogen_agentchat.agents with a local command-line executor.
    • Enables code execution tasks, specifying a working directory, timeouts, etc.

  4. Autogen Group Chat

    • Leverages RoundRobinGroupChat to let multiple agents interact in a conversation.
    • Terminates the chat based on a max number of messages or a specific terminate string.

Installation

Copy the folder

Place the ComfyUI-Autogen folder into your ComfyUI/custom_nodes/ directory.

Install dependencies

Activate the same Python environment (or virtual environment) that ComfyUI uses, then run:

pip install -r requirements.txt

Restart ComfyUI

The nodes should appear under the “Autogen” category in your ComfyUI interface.

Requirements:

autogen-core
autogen-ext
autogen-agentchat

Example

This example uses the Company Research sample from Autogen [here] Company Research Workflow *note the analyze_stock tool uses the yfinance library to retrieve stock information. It will need to be installed for this tool to work.

Todo

  • vision models need integration

Contributing

Feel free to open issues or pull requests if you run into bugs, have ideas, or want to contribute enhancements. Feedback is welcome to help improve these nodes and their capabilities.