cnnmmd

Installation: Installation in existing GUI workflow environment (ComfyUI)

You can connect your existing GUI workflow environment (ComfyUI) to this module and use it.

◯

environment

It will work from the following commit onwards: [※1]

・

ComfyUI: 2b653e8 (2025.07.11)

◯

procedure

Follow the steps below to install and configure: [※2]

・

Exclude environment construction

・

Get Custom Node

・

Copy the sample configuration files

・

Change the address of the configuration files (if necessary)

*1

It supports asynchronous node descriptions (async - await) -- in earlier versions, the execution of the node itself will result in an error.

※2

This will not work with just a custom node for ComfyUI -- the main body is the data conversion and server adjustment module on the container, and ComfyUI is the front end to it (various data flows through this module, and ComfyUI is one of the UIs that controls it) -- from ComfyUI's perspective, it is a backend that leaves heavy data conversion and tedious server adjustments to the module, allowing it to focus on generating and processing content.

Exclude environment construction

By default, this module will create a new container for the app (ComfyUI) - to prevent this container from being created, add the following plugin name to the following file:

・

${dirtop}"/cnnmmd/manage/cnf/except_custom.txt

cnnmmd_xoxxox_appcmf

*1

Since it runs on a container (as long as the ports used do not overlap), there should be no problem running it at the same time as the existing ComfyUI (if you also use it, you can flexibly change the operating state, such as port number, startup options, and CPU/GPU compatibility).

Get Custom Node

Go to the app (ComfyUI) folder and grab the custom node from the following repository:

$ cd "${dirapp}"/ComfyUI/custom_nodes
$ git clone https://github.com/cnnmmd/comfyui_xoxxox_cnnmmd

Copy the sample configuration files

Copy the following custom node configuration files to the same folder, renaming them (remove "_sample"):

$ cp -p lib/params_cmf_sample.py lib/params_cmf.py
$ cp -p lib/params_tlk_sample.py lib/params_tlk.py
$ cp -p web/lib/params_tlk_sample.js web/lib/params_tlk.js

Change the address of the configuration files (if necessary)

Specify the address and port of the intermediate connector - depending on your ComfyUI installation, this may vary:

・

lib/params_cmf.py

# If you are running it directly on your computer (host side) (including Python virtual environment and Conda virtual environment):
adrmid = "http://localhost:10001"

# If you are running in a container (Docker): If you want to share the container network with the intermediate connector:
adrmid = "http://xoxxox_appmid"

# If you are running in a container (Docker): If you do not want to share the container network with the intermediate connector:
adrmid = "http://host.docker.internal:10001"

# Other than the above (e.g. running on a different computer than the intermediate connector):
adrmid = "http://<address_host>:10001"

Specify the address and port of the web server (if you have not changed anything, the address is the same as the intermediate connector) - the way of specifying it varies depending on the installation status of ComfyUI (the web browser, which is the front end of ComfyUI, does not matter whether a container is installed or not):

・

web/lib/params_tlk.js

# If you are running it directly on your computer (host side) (including Python virtual environment and Conda virtual environment):
export const UrlLib =
'http://localhost:10002/xoxxox/lib/shared_tlk.js';
export const UrlAud =
'http://localhost:10002/xoxxox/lib/worker_aud.js';

# Other than the above (e.g. running on a different computer than the intermediate connector):
export const UrlLib =
'http://<address_host>:10002/xoxxox/lib/shared_tlk.js';
export const UrlAud =
'http://<address_host>:10002/xoxxox/lib/worker_aud.js';