A collection of custom nodes for ComfyUI that provide integration with LM Studio. These nodes allow you to load, unload, and interact with LLMs through the LM Studio command-line interface. This can be used to facilitate the automatic creation of descriptive prompts using a LLM, which is helpful for modern image models like Flux that respond well to detailed prompts in natural language.

Features

Load and manage LLM models directly from ComfyUI
Fuzzy model name searching for easy model selection
Configurable context length and parameters
Optional image model unloading to free up VRAM
Simple prompt submission interface
Option for UTF-8 safe LLM text responses
Detailed error reporting and status messages

Installation

Copy this directory to your ComfyUI/custom_nodes/ directory
Install LM Studio and make sure 'lms' command line tools are working and that 'Developer Mode' is enabled to run LM Studio as a local server.
Make sure to have at least one LLM installed inside of LM Studio.
Test this by running lms ls --detailed from the command line. You should see a list of all installed LLMs. If this doesn't work you may need to modify your PATH.
Restart ComfyUI

LM Studio

LM Studio is available at https://lmstudio.ai/
Information about LM Studio command line tools is available at https://lmstudio.ai/blog/lms

Nodes

EBU LM Studio Load Model

Unloads any loaded LLMs and then loads an LLM model into LM Studio with the following parameters:

input_string: Input text (passed through unchanged), serves to trigger the action
model_search_string: Part of the model name to search for (default: "llama"). Spaces are treated as wild cards. First matching LLM is loaded. So "nemo q5" will match "Mistral-Nemo-Instruct-2407-abliterated.Q5_K_M.gguf" for example. You can have multiple searches separated by "|" and one will be chosen at random.
context_length: Model context length (default: 4096)
seed: Random seed to ensure that this runs on each request
unload_image_models_first: Whether to unload ComfyUI image models before loading LLM

Returns:

export_string: The input string (unchanged)
model_selected: Full path of the loaded model or error message
models_found: Complete list of available models

EBU LM Studio Unload Models

Unloads all currently loaded LLM models from LM Studio.

Parameters:

input_string: Input text (passed through unchanged), serves to trigger the action
seed: Random seed to ensure that this runs on each request

Returns:

export_string: The input string (unchanged)

EBU LM Studio Unload Guider Models

Unloads all currently loaded LLM models from LM Studio, on a guider line. Great for Flux.

Parameters:

input_guider: Input text (passed through unchanged), serves to trigger the action
seed: Random seed to ensure that this runs on each request

Returns:

export_guider: The input string (unchanged)

EBU LM Studio Submit Prompt

Submits a prompt to the currently loaded LLM and returns the response.

Parameters:

prompt: The text prompt to send to the LLM
system_message: System context message (default provides image prompt generation context)
url: LM Studio API endpoint (default: http://127.0.0.1:1234/v1/chat/completions)
context_length: Context window size (default: 4096)
seed: Random seed for reproducibility
max_tokens: Maximum response length (default: 300)
temp: Temperature parameter (default: 0.5)
top_p: Top-p sampling parameter (default: 0.9)
utf8_safe_replace: Convert response to safe ASCII characters (default: false)

Returns:

generated_text: The LLM's response

EBU LM Studio Brainstormer

Generates a numbered list of ideas based on a topic, or can return a single idea chosen from a randomly generated list. Great for using an LLM to come up with names and other details for a prompt on the fly. Since LLMs tend to repeat themselves in the early part of the list from other generations I've given this the option to skip the first N results.

Parameters:

topic: The subject for brainstorming (default: "Fantasy Setting Magic Spells")
for_each_idea: Description template for each idea, without numbering (multiline)
raw_list_size: Number of ideas to generate in the full list (default: 20)
return_list_size: Number of ideas to return in the final sampled list (default: 10)
ignore_the_first: Number of initial ideas to skip when sampling (default: 0)
additional_notes_1: Optional additional note to append to the prompt (multiline)
additional_notes_2: Optional second additional note to append to the prompt (multiline)
url: LM Studio API endpoint (default: http://127.0.0.1:1234/v1/chat/completions)
context_length: Context window size (default: 4096)
seed: Random seed for reproducibility
max_tokens: Maximum response length (default: 300)
temp: Temperature parameter (default: 0.70)
top_p: Top-p sampling parameter (default: 0.95)
utf8_safe_replace: Convert response to safe ASCII characters (default: false)

Returns:

result: A formatted, numbered list of sampled ideas
full_list: The complete list of generated ideas with original numbering

Requirements

Python 3.x
ComfyUI
LM Studio with CLI access and local server configured

Common Issues and Solutions

"Can't connect to LM Studio"

Ensure LM Studio is running
Check that LMS CLI Access is healthy
Verify the port number matches your LM Studio settings
Confirm LM Studio is in your system PATH

"no model by that name found"

Check that the model name search string matches part of your model's name
Use lms ls --detailed in terminal to see available models
Models must be added to LM Studio before they can be loaded

Model Loading Issues

Ensure you have enough VRAM available
Try unloading image models first
Check LM Studio logs for detailed error messages

License

This project is licensed under the MIT License - see the LICENSE file for details

Acknowledgments

Initially inspired by the LLM node from CrasH Utils Custom Nodes
Thanks to the ComfyUI and LM Studio teams

For more help

See the included example workflow flux_example_ebu_lmstudio.json to see how these three nodes can work together for VRAM efficient prompt generation to image generation with Flux.